You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@openoffice.apache.org by Peter Kovacs <pe...@apache.org> on 2020/12/07 18:38:15 UTC
[discussion] future embedded DB in AOO
Hi all,
I would like to know how you guys feel we should move on with our base
Database. Our current strategy is a small sized embedded DB using HSQLDB
(BSD). There are 2 other options in the same weight class, that is H2
(EPL1 or MPL2) and Apache Derby (AL2).
Something like SQLLite (PublicDomain) could be technical interesting,
but I think their licensing is not so appealing.
We could also consider firebird (MPL) like LO, but I have the impression
that this is not a lightweight DB, but featurerich. And I wonder if this
is really something we should provide. I would rather develop a way to
make it easy to switch the DB when a Project grows.
What are your thoughts?
All the Best
Peter
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
RE: [discussion] future embedded DB in AOO
Posted by Jörg Schmidt <jo...@j-m-schmidt.de>.
> -----Original Message-----
> From: Matthias Seidel [mailto:matthias.seidel@hamburg.de]
> Sent: Tuesday, December 08, 2020 11:15 AM
> To: dev@openoffice.apache.org
> Subject: Re: [discussion] future embedded DB in AOO
>
> Hi Peter,
>
> If we switch to some other embedded database engine we need
> to provide a
> way to let users open their existing databases.
>
> According to the German Wikipedia article [1] H2 provides a
> "compatibility mode" for that purpose.
>
> Personally I don't see any reason to move away from Java.
If that is so, that is so. But what changed your mind _or did I misrepresent your earlier statements about licensing problems_?
(OpenJDK was already a topic back then, but if I remember correctly your opinion was rather reserved.)
Jörg
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: [discussion] future embedded DB in AOO
Posted by Matthias Seidel <ma...@hamburg.de>.
Hi Peter,
If we switch to some other embedded database engine we need to provide a
way to let users open their existing databases.
According to the German Wikipedia article [1] H2 provides a
"compatibility mode" for that purpose.
Personally I don't see any reason to move away from Java.
Regards,
Matthias
[1] https://de.wikipedia.org/wiki/H2_Database#Funktionen
Am 08.12.20 um 10:14 schrieb Peter Kovacs:
>
> On 08.12.20 09:27, Jörg Schmidt wrote:
>>> -----Original Message-----
>>> From: Keith N. McKenna [mailto:keith.mckenna@comcast.net]
>>> Sent: Tuesday, December 08, 2020 3:49 AM
>>> To: dev@openoffice.apache.org
>>> Subject: Re: [discussion] future embedded DB in AOO
>>> My first question has to be : Why are we still using such an
>>> old version
>>> of HSQLDB? As memory serves it is `1.8.0 and the latest release of
>>> HSQLDB is 2.5.1.
>
> I thought we had discussion on this, but I could not find them.
>
> I will have later a second look. But this is a good thought. How much
> time will an upgrade bring us?
>
> It will influence the roadmap for sure.
>
>>>
>>> Second is why cannot we upgrade to HSQLDB 2.X.
>
> The question is not if we can, the question is if it is worth it.
>
> I would like to take an early opportunity to look beyond the immediate
> need of keep things working.
>
> I put on the question what alternative exists. Jörg brought up the
> Question how much Java do we want in future?
>
>>> As a process engineer and not a programmer these are simple
>>> questions I
>>> need answers for to intelligently contribute to this
>>> discussion. Thanks
>>> all for your understanding.
>
> All opinion and thoughts are welcome. I want to find out where we want
> to head with the embedded DB. All Ideas and reasoning can bring up
> good Views.
>
> We can also discuss where we want to go with Base. The question is
> Connected. And Base needs some love, don't you think?
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
> For additional commands, e-mail: dev-help@openoffice.apache.org
>
Re: [discussion] future embedded DB in AOO
Posted by Peter Kovacs <pe...@apache.org>.
On 08.12.20 09:27, Jörg Schmidt wrote:
>> -----Original Message-----
>> From: Keith N. McKenna [mailto:keith.mckenna@comcast.net]
>> Sent: Tuesday, December 08, 2020 3:49 AM
>> To: dev@openoffice.apache.org
>> Subject: Re: [discussion] future embedded DB in AOO
>> My first question has to be : Why are we still using such an
>> old version
>> of HSQLDB? As memory serves it is `1.8.0 and the latest release of
>> HSQLDB is 2.5.1.
I thought we had discussion on this, but I could not find them.
I will have later a second look. But this is a good thought. How much
time will an upgrade bring us?
It will influence the roadmap for sure.
>>
>> Second is why cannot we upgrade to HSQLDB 2.X.
The question is not if we can, the question is if it is worth it.
I would like to take an early opportunity to look beyond the immediate
need of keep things working.
I put on the question what alternative exists. Jörg brought up the
Question how much Java do we want in future?
>> As a process engineer and not a programmer these are simple
>> questions I
>> need answers for to intelligently contribute to this
>> discussion. Thanks
>> all for your understanding.
All opinion and thoughts are welcome. I want to find out where we want
to head with the embedded DB. All Ideas and reasoning can bring up good
Views.
We can also discuss where we want to go with Base. The question is
Connected. And Base needs some love, don't you think?
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
RE: [discussion] future embedded DB in AOO
Posted by Jörg Schmidt <jo...@j-m-schmidt.de>.
> -----Original Message-----
> From: Keith N. McKenna [mailto:keith.mckenna@comcast.net]
> Sent: Tuesday, December 08, 2020 3:49 AM
> To: dev@openoffice.apache.org
> Subject: Re: [discussion] future embedded DB in AOO
> My first question has to be : Why are we still using such an
> old version
> of HSQLDB? As memory serves it is `1.8.0 and the latest release of
> HSQLDB is 2.5.1.
>
> Second is why cannot we upgrade to HSQLDB 2.X.
>
> As a process engineer and not a programmer these are simple
> questions I
> need answers for to intelligently contribute to this
> discussion. Thanks
> all for your understanding.
indeed, one should not leave out these questions.
Jörg
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: [discussion] future embedded DB in AOO
Posted by Mechtilde <oo...@mechtilde.de>.
Hello Keith,
as i know from former time (before comming to apache)
we evaluate that we can't update HSQL Db to version 2.x. that is the
reason, why also Debian shipped both version of HSQLDB.
Kind regards
Mechtilde
Am 08.12.20 um 03:48 schrieb Keith N. McKenna:
> On 12/7/2020 1:38 PM, Peter Kovacs wrote:
>> Hi all,
>>
>> I would like to know how you guys feel we should move on with our base
>> Database. Our current strategy is a small sized embedded DB using HSQLDB
>> (BSD). There are 2 other options in the same weight class, that is H2
>> (EPL1 or MPL2) and Apache Derby (AL2).
>>
>> Something like SQLLite (PublicDomain) could be technical interesting,
>> but I think their licensing is not so appealing.
>>
>> We could also consider firebird (MPL) like LO, but I have the impression
>> that this is not a lightweight DB, but featurerich. And I wonder if this
>> is really something we should provide. I would rather develop a way to
>> make it easy to switch the DB when a Project grows.
>>
>> What are your thoughts?
>>
>>
>> All the Best
>>
>> Peter
>
> My first question has to be : Why are we still using such an old version
> of HSQLDB? As memory serves it is `1.8.0 and the latest release of
> HSQLDB is 2.5.1.
>
> Second is why cannot we upgrade to HSQLDB 2.X.
>
> As a process engineer and not a programmer these are simple questions I
> need answers for to intelligently contribute to this discussion. Thanks
> all for your understanding.
>
> Regards
> Keith.
>
--
Mechtilde Stehmann
## Apache OpenOffice
## Freie Office Suite für Linux, MacOSX, Windows und OS/2
## Debian Developer
## PGP encryption welcome
## F0E3 7F3D C87A 4998 2899 39E7 F287 7BBA 141A AD7F
Re: [discussion] future embedded DB in AOO
Posted by "Keith N. McKenna" <ke...@comcast.net>.
On 12/7/2020 1:38 PM, Peter Kovacs wrote:
> Hi all,
>
> I would like to know how you guys feel we should move on with our base
> Database. Our current strategy is a small sized embedded DB using HSQLDB
> (BSD). There are 2 other options in the same weight class, that is H2
> (EPL1 or MPL2) and Apache Derby (AL2).
>
> Something like SQLLite (PublicDomain) could be technical interesting,
> but I think their licensing is not so appealing.
>
> We could also consider firebird (MPL) like LO, but I have the impression
> that this is not a lightweight DB, but featurerich. And I wonder if this
> is really something we should provide. I would rather develop a way to
> make it easy to switch the DB when a Project grows.
>
> What are your thoughts?
>
>
> All the Best
>
> Peter
My first question has to be : Why are we still using such an old version
of HSQLDB? As memory serves it is `1.8.0 and the latest release of
HSQLDB is 2.5.1.
Second is why cannot we upgrade to HSQLDB 2.X.
As a process engineer and not a programmer these are simple questions I
need answers for to intelligently contribute to this discussion. Thanks
all for your understanding.
Regards
Keith.
Re: [discussion] future embedded DB in AOO
Posted by Peter Kovacs <pe...@apache.org>.
On 07.12.20 23:03, Jim Jagielski wrote:
>
>> On Dec 7, 2020, at 4:27 PM, Peter Kovacs <pe...@Apache.org> wrote:
>>
>>
>> On 07.12.20 20:48, Jim Jagielski wrote:
>>>> On Dec 7, 2020, at 1:38 PM, Peter Kovacs <pe...@Apache.org> wrote:
>>>>
>>>> Hi all,
>>>>
>>>> I would like to know how you guys feel we should move on with our base Database. Our current strategy is a small sized embedded DB using HSQLDB (BSD). There are 2 other options in the same weight class, that is H2 (EPL1 or MPL2) and Apache Derby (AL2).
>>>>
>>>> Something like SQLLite (PublicDomain) could be technical interesting, but I think their licensing is not so appealing.
>>> Is EPL1/MPL2 more appealing than SQLite's PD?
>> No not really. I am intimidated by the statements:
>>
>> Open-Source, not Open-Contribution
>>
>> SQLite is open-source, meaning that you can make as many copies of
>> it as you want and do whatever you want with those copies, without
>> limitation. But SQLite is not open-contribution. In order to keep
>> SQLite in the public domain and ensure that the code does not become
>> contaminated with proprietary or licensed content, the project does
>> not accept patches from unknown persons.
>>
>> All of the code in SQLite is original, having been written
>> specifically for use by SQLite. No code has been copied from unknown
>> sources on the internet.
>>
>> Contributed Code
>>
>> In order to keep SQLite completely free and unencumbered by
>> copyright, the project does not accept patches. If you would like to
>> make a suggested change, and include a patch as a proof-of-concept,
>> that would be great. However please do not be offended if we rewrite
>> your patch from scratch.
>>
>> Actually their Licensing is:
>>
>> Anyone is free to copy, modify, publish, use, compile, sell, or
>> distribute the original SQLite code, either in source code form or
>> as a compiled binary, for any purpose, commercial or non-commercial,
>> and by any means.
>>
>> Which is fine and works for us.
> Also, let's look at https://www.apache.org/legal/resolved.html
>
> Works in the public domain (or covered by a license treated similarly) may be included within Apache products. Attribution is required (in a similar fashion to the Category A list.
>
> A work should be treated as being in the public domain when one of the following applies:
>
> the work is covered by
> the Creative Commons Public Domain Mark <http://creativecommons.org/publicdomain/mark/1.0/>, or
> a suitable dedication (to the public domain) by the authors; or
>
>
> I think we can confidently affirm the 2nd bullet.
thanks Jim. I was not that sure about this point.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: [discussion] future embedded DB in AOO
Posted by Jim Jagielski <ji...@jaguNET.com>.
> On Dec 7, 2020, at 4:27 PM, Peter Kovacs <pe...@Apache.org> wrote:
>
>
> On 07.12.20 20:48, Jim Jagielski wrote:
>>
>>> On Dec 7, 2020, at 1:38 PM, Peter Kovacs <pe...@Apache.org> wrote:
>>>
>>> Hi all,
>>>
>>> I would like to know how you guys feel we should move on with our base Database. Our current strategy is a small sized embedded DB using HSQLDB (BSD). There are 2 other options in the same weight class, that is H2 (EPL1 or MPL2) and Apache Derby (AL2).
>>>
>>> Something like SQLLite (PublicDomain) could be technical interesting, but I think their licensing is not so appealing.
>> Is EPL1/MPL2 more appealing than SQLite's PD?
>
> No not really. I am intimidated by the statements:
>
> Open-Source, not Open-Contribution
>
> SQLite is open-source, meaning that you can make as many copies of
> it as you want and do whatever you want with those copies, without
> limitation. But SQLite is not open-contribution. In order to keep
> SQLite in the public domain and ensure that the code does not become
> contaminated with proprietary or licensed content, the project does
> not accept patches from unknown persons.
>
> All of the code in SQLite is original, having been written
> specifically for use by SQLite. No code has been copied from unknown
> sources on the internet.
>
> Contributed Code
>
> In order to keep SQLite completely free and unencumbered by
> copyright, the project does not accept patches. If you would like to
> make a suggested change, and include a patch as a proof-of-concept,
> that would be great. However please do not be offended if we rewrite
> your patch from scratch.
>
> Actually their Licensing is:
>
> Anyone is free to copy, modify, publish, use, compile, sell, or
> distribute the original SQLite code, either in source code form or
> as a compiled binary, for any purpose, commercial or non-commercial,
> and by any means.
>
> Which is fine and works for us.
Also, let's look at https://www.apache.org/legal/resolved.html
Works in the public domain (or covered by a license treated similarly) may be included within Apache products. Attribution is required (in a similar fashion to the Category A list.
A work should be treated as being in the public domain when one of the following applies:
the work is covered by
the Creative Commons Public Domain Mark <http://creativecommons.org/publicdomain/mark/1.0/>, or
a suitable dedication (to the public domain) by the authors; or
I think we can confidently affirm the 2nd bullet.
Re: [discussion] future embedded DB in AOO
Posted by Peter Kovacs <pe...@apache.org>.
On 07.12.20 20:48, Jim Jagielski wrote:
>
>> On Dec 7, 2020, at 1:38 PM, Peter Kovacs <pe...@Apache.org> wrote:
>>
>> Hi all,
>>
>> I would like to know how you guys feel we should move on with our base Database. Our current strategy is a small sized embedded DB using HSQLDB (BSD). There are 2 other options in the same weight class, that is H2 (EPL1 or MPL2) and Apache Derby (AL2).
>>
>> Something like SQLLite (PublicDomain) could be technical interesting, but I think their licensing is not so appealing.
> Is EPL1/MPL2 more appealing than SQLite's PD?
No not really. I am intimidated by the statements:
Open-Source, not Open-Contribution
SQLite is open-source, meaning that you can make as many copies of
it as you want and do whatever you want with those copies, without
limitation. But SQLite is not open-contribution. In order to keep
SQLite in the public domain and ensure that the code does not become
contaminated with proprietary or licensed content, the project does
not accept patches from unknown persons.
All of the code in SQLite is original, having been written
specifically for use by SQLite. No code has been copied from unknown
sources on the internet.
Contributed Code
In order to keep SQLite completely free and unencumbered by
copyright, the project does not accept patches. If you would like to
make a suggested change, and include a patch as a proof-of-concept,
that would be great. However please do not be offended if we rewrite
your patch from scratch.
Actually their Licensing is:
Anyone is free to copy, modify, publish, use, compile, sell, or
distribute the original SQLite code, either in source code form or
as a compiled binary, for any purpose, commercial or non-commercial,
and by any means.
Which is fine and works for us.
Re: [discussion] future embedded DB in AOO
Posted by Jim Jagielski <ji...@jaguNET.com>.
> On Dec 7, 2020, at 1:38 PM, Peter Kovacs <pe...@Apache.org> wrote:
>
> Hi all,
>
> I would like to know how you guys feel we should move on with our base Database. Our current strategy is a small sized embedded DB using HSQLDB (BSD). There are 2 other options in the same weight class, that is H2 (EPL1 or MPL2) and Apache Derby (AL2).
>
> Something like SQLLite (PublicDomain) could be technical interesting, but I think their licensing is not so appealing.
Is EPL1/MPL2 more appealing than SQLite's PD?
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
RE: [discussion] future embedded DB in AOO
Posted by Jörg Schmidt <jo...@j-m-schmidt.de>.
Hello Mechtilde,
> -----Original Message-----
> From: Mechtilde [mailto:ooo@mechtilde.de]
> Sent: Tuesday, December 08, 2020 2:08 PM
> To: dev@openoffice.apache.org
> Subject: Re: [discussion] future embedded DB in AOO
>
> Hello Jörg,
>
> Am 08.12.20 um 11:29 schrieb Jörg Schmidt:
> >
> >
> >> -----Original Message-----
> >> From: Mechtilde [mailto:ooo@mechtilde.de]
> >> Sent: Tuesday, December 08, 2020 10:34 AM
> >> To: dev@openoffice.apache.org
> >> Subject: Re: [discussion] future embedded DB in AOO
> >
> >>> As far as I know from Mathias Java is, in the long run, a
> >> problem because of the licensing policy of Oracle(*) and we
> >> have to find ways to free ourselves from this dependency.
> >>
> >> This isn't a problem anymore, because we can use Adopt
> >> OpenJDK, which is
> >> used in the linux distributions since long time.
> >
> > I know that OpenJDK can be used as an alternative in Linux,
> but what is the situation in Windows? I have no experience
> with OpenJDK in Windows.
> >
>
> this is also available for Windows
That is clear, but what are the practical experiences?
Excuse me, I was the only one who was made aware of the Java problem by Mathias (about 1.5 years ago?) and at that time the OpenJDK seemed to me to be a little bit critical (in Mathias' evaluation but also in third parties' evaluation).
It would be/is of course great if there were no more considerations, just in a discussion that Peter also considers strategic, one should question such things as a precaution.
That is all I am interested in.
greetings,
Jörg
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: [discussion] future embedded DB in AOO
Posted by Rory O'Farrell <of...@iol.ie>.
On Tue, 8 Dec 2020 14:26:45 +0100
Matthias Seidel <ma...@hamburg.de> wrote:
> Hi Mechtilde,
>
> Am 08.12.20 um 14:07 schrieb Mechtilde:
> > Hello Jörg,
> >
> > Am 08.12.20 um 11:29 schrieb Jörg Schmidt:
> >>
> >>
> >>> -----Original Message-----
> >>> From: Mechtilde [mailto:ooo@mechtilde.de]
> >>> Sent: Tuesday, December 08, 2020 10:34 AM
> >>> To: dev@openoffice.apache.org
> >>> Subject: Re: [discussion] future embedded DB in AOO
> >>>> As far as I know from Mathias Java is, in the long run, a
> >>> problem because of the licensing policy of Oracle(*) and we
> >>> have to find ways to free ourselves from this dependency.
> >>>
> >>> This isn't a problem anymore, because we can use Adopt
> >>> OpenJDK, which is
> >>> used in the linux distributions since long time.
> >> I know that OpenJDK can be used as an alternative in Linux, but what is the situation in Windows? I have no experience with OpenJDK in Windows.
> >>
> > this is also available for Windows
>
> Indeed, all my Windows test build are done with AdoptOpenJDK for quite
> some time now and we officially support AdoptOpenJDK (on every platform)
> since AOO 4.1.7 [1]
>
> Regards,
>
> Matthias
>
> [1]
> https://cwiki.apache.org/confluence/display/OOOUSERS/AOO+4.1.7+Release+Notes#AOO4.1.7ReleaseNotes-Improvements/Enhancements
>
> >
> >> Jörg
> >
>
Have used AdoptOpenJDK on Xubuntu towards end of 18:04 and since launch of 20:04 on several computers with no problem.
--
Rory O'Farrell <of...@iol.ie>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: [discussion] future embedded DB in AOO
Posted by Matthias Seidel <ma...@hamburg.de>.
Hi Mechtilde,
Am 08.12.20 um 14:07 schrieb Mechtilde:
> Hello Jörg,
>
> Am 08.12.20 um 11:29 schrieb Jörg Schmidt:
>>
>>
>>> -----Original Message-----
>>> From: Mechtilde [mailto:ooo@mechtilde.de]
>>> Sent: Tuesday, December 08, 2020 10:34 AM
>>> To: dev@openoffice.apache.org
>>> Subject: Re: [discussion] future embedded DB in AOO
>>>> As far as I know from Mathias Java is, in the long run, a
>>> problem because of the licensing policy of Oracle(*) and we
>>> have to find ways to free ourselves from this dependency.
>>>
>>> This isn't a problem anymore, because we can use Adopt
>>> OpenJDK, which is
>>> used in the linux distributions since long time.
>> I know that OpenJDK can be used as an alternative in Linux, but what is the situation in Windows? I have no experience with OpenJDK in Windows.
>>
> this is also available for Windows
Indeed, all my Windows test build are done with AdoptOpenJDK for quite
some time now and we officially support AdoptOpenJDK (on every platform)
since AOO 4.1.7 [1]
Regards,
Matthias
[1]
https://cwiki.apache.org/confluence/display/OOOUSERS/AOO+4.1.7+Release+Notes#AOO4.1.7ReleaseNotes-Improvements/Enhancements
>
>> Jörg
>
Re: [discussion] future embedded DB in AOO
Posted by Mechtilde <oo...@mechtilde.de>.
Hello Jörg,
Am 08.12.20 um 11:29 schrieb Jörg Schmidt:
>
>
>> -----Original Message-----
>> From: Mechtilde [mailto:ooo@mechtilde.de]
>> Sent: Tuesday, December 08, 2020 10:34 AM
>> To: dev@openoffice.apache.org
>> Subject: Re: [discussion] future embedded DB in AOO
>
>>> As far as I know from Mathias Java is, in the long run, a
>> problem because of the licensing policy of Oracle(*) and we
>> have to find ways to free ourselves from this dependency.
>>
>> This isn't a problem anymore, because we can use Adopt
>> OpenJDK, which is
>> used in the linux distributions since long time.
>
> I know that OpenJDK can be used as an alternative in Linux, but what is the situation in Windows? I have no experience with OpenJDK in Windows.
>
this is also available for Windows
>
> Jörg
--
Mechtilde Stehmann
## Apache OpenOffice
## Freie Office Suite für Linux, MacOSX, Windows und OS/2
## Debian Developer
## PGP encryption welcome
## F0E3 7F3D C87A 4998 2899 39E7 F287 7BBA 141A AD7F
RE: [discussion] future embedded DB in AOO
Posted by Jörg Schmidt <jo...@j-m-schmidt.de>.
> -----Original Message-----
> From: Mechtilde [mailto:ooo@mechtilde.de]
> Sent: Tuesday, December 08, 2020 10:34 AM
> To: dev@openoffice.apache.org
> Subject: Re: [discussion] future embedded DB in AOO
> > As far as I know from Mathias Java is, in the long run, a
> problem because of the licensing policy of Oracle(*) and we
> have to find ways to free ourselves from this dependency.
>
> This isn't a problem anymore, because we can use Adopt
> OpenJDK, which is
> used in the linux distributions since long time.
I know that OpenJDK can be used as an alternative in Linux, but what is the situation in Windows? I have no experience with OpenJDK in Windows.
Jörg
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: [discussion] future embedded DB in AOO
Posted by Mechtilde <oo...@mechtilde.de>.
Hello
onle one statement inside
Am 08.12.20 um 10:23 schrieb Jörg Schmidt:
>
>
>> -----Original Message-----
>> From: Peter Kovacs [mailto:petko@apache.org]
>> Sent: Tuesday, December 08, 2020 9:46 AM
>> To: dev@openoffice.apache.org
>> Subject: Re: [discussion] future embedded DB in AOO
>>
>>
>> On 08.12.20 09:14, Jörg Schmidt wrote:
>>>
>>>
>>>> -----Original Message-----
>>>> From: Peter Kovacs [mailto:petko@apache.org]
>>>> Sent: Monday, December 07, 2020 11:03 PM
>>>> To: dev@openoffice.apache.org
>>>> Subject: Re: [discussion] future embedded DB in AOO
>>>>> Maybe one more thing: Is the problem database even important?
>>>> We cannot build with Java 11 and up. Decisions have to be
>> made soon.
>>>> Strategy has to be setup. A Roadmap would be great
>>> Peter, I do not understand what that means in concrete terms.
>>>
>>> IF Java is a critical factor (I guess because of Oracle's
>> licensing policy) why is it less so for H2 or Derby than for HSQLDB?
>>
>> No, I meant that building OpenOffice with Java 11 and later
>> causes the
>> build to break on hsqldb. So we need to talk
>>
>> what we do. Current state is that we need to change the code
>> in order to
>> build on a later Java Version. That is what brought this
>> topic up. And
>> again I want to have a discussion on what we head out for. What is
>> needed. And I think this makes it a good time to collect all
>> opinions,
>> arguments and requirements.
>
> Thank you for this declaration. The situation is therefore understandable to me.
>
>>> Actually we should rather choose a database that works
>> without Java(?)
>>>
>>> Also: IF the Java question is urgent, why is it only for
>> Base? Doesn't belong then on the agenda anymore?
>>
>> Where do I say it is urgent?
>
> I had understood "We cannot build with Java 11 and up." that way.
>
>> This takes time. It is a good time to start this now.
>
> yes, this is true
>
>
>> What is your opinion on Java? Not that I want to start a general Java
>> discussion, but you are right they are connected.
>
> As far as I know from Mathias Java is, in the long run, a problem because of the licensing policy of Oracle(*) and we have to find ways to free ourselves from this dependency.
This isn't a problem anymore, because we can use Adopt OpenJDK, which is
used in the linux distributions since long time.
>
> Since these are rather long-term considerations, I can't say if we have to consider this already now for a decision regarding Base.
> It would tend to make sense, but searching for a database without Java dependency might collide somewhat with the question of multiplatform capability.
>
> But please ... I am only answering your question here, I do not want to encourage you to expand the discussion.
>
>
> (*)
> if I understood it correctly, the problem is not technical in essence, but future Java versions should be subject to license fees, at least for business use (?)
>
> Jörg
>
--
Mechtilde Stehmann
## Apache OpenOffice
## Freie Office Suite für Linux, MacOSX, Windows und OS/2
## Debian Developer
## PGP encryption welcome
## F0E3 7F3D C87A 4998 2899 39E7 F287 7BBA 141A AD7F
RE: [discussion] future embedded DB in AOO
Posted by Jörg Schmidt <jo...@j-m-schmidt.de>.
> -----Original Message-----
> From: Peter Kovacs [mailto:petko@apache.org]
> Sent: Tuesday, December 08, 2020 9:46 AM
> To: dev@openoffice.apache.org
> Subject: Re: [discussion] future embedded DB in AOO
>
>
> On 08.12.20 09:14, Jörg Schmidt wrote:
> >
> >
> >> -----Original Message-----
> >> From: Peter Kovacs [mailto:petko@apache.org]
> >> Sent: Monday, December 07, 2020 11:03 PM
> >> To: dev@openoffice.apache.org
> >> Subject: Re: [discussion] future embedded DB in AOO
> >>> Maybe one more thing: Is the problem database even important?
> >> We cannot build with Java 11 and up. Decisions have to be
> made soon.
> >> Strategy has to be setup. A Roadmap would be great
> > Peter, I do not understand what that means in concrete terms.
> >
> > IF Java is a critical factor (I guess because of Oracle's
> licensing policy) why is it less so for H2 or Derby than for HSQLDB?
>
> No, I meant that building OpenOffice with Java 11 and later
> causes the
> build to break on hsqldb. So we need to talk
>
> what we do. Current state is that we need to change the code
> in order to
> build on a later Java Version. That is what brought this
> topic up. And
> again I want to have a discussion on what we head out for. What is
> needed. And I think this makes it a good time to collect all
> opinions,
> arguments and requirements.
Thank you for this declaration. The situation is therefore understandable to me.
> > Actually we should rather choose a database that works
> without Java(?)
> >
> > Also: IF the Java question is urgent, why is it only for
> Base? Doesn't belong then on the agenda anymore?
>
> Where do I say it is urgent?
I had understood "We cannot build with Java 11 and up." that way.
> This takes time. It is a good time to start this now.
yes, this is true
> What is your opinion on Java? Not that I want to start a general Java
> discussion, but you are right they are connected.
As far as I know from Mathias Java is, in the long run, a problem because of the licensing policy of Oracle(*) and we have to find ways to free ourselves from this dependency.
Since these are rather long-term considerations, I can't say if we have to consider this already now for a decision regarding Base.
It would tend to make sense, but searching for a database without Java dependency might collide somewhat with the question of multiplatform capability.
But please ... I am only answering your question here, I do not want to encourage you to expand the discussion.
(*)
if I understood it correctly, the problem is not technical in essence, but future Java versions should be subject to license fees, at least for business use (?)
Jörg
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: [discussion] future embedded DB in AOO
Posted by Peter Kovacs <pe...@apache.org>.
On 08.12.20 09:14, Jörg Schmidt wrote:
>
>
>> -----Original Message-----
>> From: Peter Kovacs [mailto:petko@apache.org]
>> Sent: Monday, December 07, 2020 11:03 PM
>> To: dev@openoffice.apache.org
>> Subject: Re: [discussion] future embedded DB in AOO
>>> Maybe one more thing: Is the problem database even important?
>> We cannot build with Java 11 and up. Decisions have to be made soon.
>> Strategy has to be setup. A Roadmap would be great
> Peter, I do not understand what that means in concrete terms.
>
> IF Java is a critical factor (I guess because of Oracle's licensing policy) why is it less so for H2 or Derby than for HSQLDB?
No, I meant that building OpenOffice with Java 11 and later causes the
build to break on hsqldb. So we need to talk
what we do. Current state is that we need to change the code in order to
build on a later Java Version. That is what brought this topic up. And
again I want to have a discussion on what we head out for. What is
needed. And I think this makes it a good time to collect all opinions,
arguments and requirements.
> Actually we should rather choose a database that works without Java(?)
>
> Also: IF the Java question is urgent, why is it only for Base? Doesn't belong then on the agenda anymore?
Where do I say it is urgent? I say we should take some time and consider
our options. Making a sound decision.
This takes time. It is a good time to start this now.
>
>> I want to take
>> some time,
>> develop some strategy Maybe a roadmap.
> IF it's a strategy we should probably talk about Java and not only about Base.
What is your opinion on Java? Not that I want to start a general Java
discussion, but you are right they are connected.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
RE: [discussion] future embedded DB in AOO
Posted by Jörg Schmidt <jo...@j-m-schmidt.de>.
> -----Original Message-----
> From: Jörg Schmidt [mailto:joesch@j-m-schmidt.de]
> Sent: Tuesday, December 08, 2020 9:14 AM
> To: dev@openoffice.apache.org
> Subject: RE: [discussion] future embedded DB in AOO
>
>
> [...]
Now I have to focus again:
Is the concrete problem Base or Java?
IF it is Java why are we even talking about replacing HSQLDB with Java-based databases?
Shouldn't our goal then be to find a database that is not based on Java?
Jörg
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
RE: [discussion] future embedded DB in AOO
Posted by Jörg Schmidt <jo...@j-m-schmidt.de>.
> -----Original Message-----
> From: Peter Kovacs [mailto:petko@apache.org]
> Sent: Monday, December 07, 2020 11:03 PM
> To: dev@openoffice.apache.org
> Subject: Re: [discussion] future embedded DB in AOO
> > Maybe one more thing: Is the problem database even important?
> We cannot build with Java 11 and up. Decisions have to be made soon.
> Strategy has to be setup. A Roadmap would be great
Peter, I do not understand what that means in concrete terms.
IF Java is a critical factor (I guess because of Oracle's licensing policy) why is it less so for H2 or Derby than for HSQLDB?
Actually we should rather choose a database that works without Java(?)
Also: IF the Java question is urgent, why is it only for Base? Doesn't belong then on the agenda anymore?
> I want to take
> some time,
> develop some strategy Maybe a roadmap.
IF it's a strategy we should probably talk about Java and not only about Base.
greetings,
Jörg
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: [discussion] future embedded DB in AOO
Posted by Peter Kovacs <pe...@apache.org>.
On 07.12.20 22:35, Jörg Schmidt wrote:
>> -----Original Message-----
>> From: Jim Jagielski [mailto:jim@jaguNET.com]
>> Sent: Monday, December 07, 2020 10:11 PM
>> To: dev
>> Subject: Re: [discussion] future embedded DB in AOO
>>
>> I agree that eating our own dogfood and using Derby would be
>> the "better" path to take; plus it is relatively small while
>> at the same time being quite powerful and compliant.
> As already written: I don't know both systems and I only made my statement based on the information I found on Wikipedia.
> If Mechtilde (and others) have real experience with one of the two databases, this would have more weight.
>
> Maybe one more thing: Is the problem database even important?
We cannot build with Java 11 and up. Decisions have to be made soon.
Strategy has to be setup. A Roadmap would be great
> LO, for example, has much more manpower than we do, and yet Firebird in LO still seems to me to be a construction site, because actually everyone advises me to stay with HSQLDB (as far as an embedded DB is concerned).
More Manpower does not mean they invest it into Base. So I would not
judge on our strength or weakness.
> AOO should avoid getting into the same situation, because invested work that ultimately creates no benefit for the user is badly invested.
That statement is true but incomplete. We need to take into account how
much work it is for us to maintain an embedded DB. The DB eats our
resources. A good choice can make a difference in the future.
> Yes, it is a very soft argument to warn not to start at all, because you are afraid, that you can not make it.
> But many other questions seem to me to be more urgent, e.g. the currently detected problems with Big Sur (maybe other MacOS versions?)
Well, there are always more important topics. I want to take some time,
develop some strategy Maybe a roadmap. Maybe it is a good Idea to check
the options a bit. Maybe even to have something to talk about on next
FOSDEM.
We have a Room and a Deadline and no talks yet. :)
I just want to know how everyone feels on this topic, since we need to
make a decision someday.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
RE: [discussion] future embedded DB in AOO
Posted by Jörg Schmidt <jo...@j-m-schmidt.de>.
> -----Original Message-----
> From: Jim Jagielski [mailto:jim@jaguNET.com]
> Sent: Monday, December 07, 2020 10:11 PM
> To: dev
> Subject: Re: [discussion] future embedded DB in AOO
>
> I agree that eating our own dogfood and using Derby would be
> the "better" path to take; plus it is relatively small while
> at the same time being quite powerful and compliant.
As already written: I don't know both systems and I only made my statement based on the information I found on Wikipedia.
If Mechtilde (and others) have real experience with one of the two databases, this would have more weight.
Maybe one more thing: Is the problem database even important?
LO, for example, has much more manpower than we do, and yet Firebird in LO still seems to me to be a construction site, because actually everyone advises me to stay with HSQLDB (as far as an embedded DB is concerned).
AOO should avoid getting into the same situation, because invested work that ultimately creates no benefit for the user is badly invested.
Yes, it is a very soft argument to warn not to start at all, because you are afraid, that you can not make it.
But many other questions seem to me to be more urgent, e.g. the currently detected problems with Big Sur (maybe other MacOS versions?)
Jörg
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: [discussion] future embedded DB in AOO
Posted by Carl Marcum <cm...@apache.org>.
On 12/7/20 4:11 PM, Jim Jagielski wrote:
> I agree that eating our own dogfood and using Derby would be the "better" path to take; plus it is relatively small while at the same time being quite powerful and compliant.
I think that is a good case for Derby over the others unless one has
compelling features not offered.
>
>> On Dec 7, 2020, at 2:46 PM, Jörg Schmidt <jo...@j-m-schmidt.de> wrote:
>>
>> Hello,
>>
>>> -----Original Message-----
>>> From: Peter Kovacs [mailto:petko@apache.org]
>>> Sent: Monday, December 07, 2020 7:38 PM
>>> To: dev@openoffice.apache.org
>>> Subject: [discussion] future embedded DB in AOO
>>>
>>> Hi all,
>>>
>>> I would like to know how you guys feel we should move on with
>>> our base
>>> Database. ...
>> There is no compelling reason for us to join LO on the database issue, it may even be strategically smart not to.
>>
>> I know neither H2 nor Apache Derby and have only briefly checked Wikipedia and if these two were available I would be in favour of Derby, because:
>> -obviously better support for SQL
>> -maybe, due to the longer history, the more mature system
>> -apparently better encryption features
>> -Apache Project
>>
>> But maybe there is an argument for H2:
>> this project seems to be, in essence, driven by a single developer, most likely the one who would welcome H2 to become more known when we use it in AOO and would be willing to coordinate the development of H2 with us, so that H2 adapts to the requirements of AOO in the best possible way.
>>
>>
>> With both databases, however, Java bothers me - or have we given up the intention of making AOO less dependent on Java?
>> (yes, of course this is a very long-term question)
>>
>>
>>
>>
>> greetings,
>> Jörg
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org <ma...@openoffice.apache.org>
>> For additional commands, e-mail: dev-help@openoffice.apache.org <ma...@openoffice.apache.org>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: [discussion] future embedded DB in AOO
Posted by Mechtilde <oo...@mechtilde.de>.
Hello Carl
Am 08.12.20 um 13:56 schrieb Carl Marcum:
> Hi All,
>
> On 12/7/20 4:11 PM, Jim Jagielski wrote:
>> I agree that eating our own dogfood and using Derby would be the
>> "better" path to take; plus it is relatively small while at the same
>> time being quite powerful and compliant.
>
> It's not well known but Derby is also rebranded and included in the JDK
> as JavaDB.
>
> I have used Derby in Embedded and also Client/Server modes and found it
> to be a good solution.
>
> I do a lot of Grails development and Grails uses H2 in development mode
> and it has a nice web console available.
> I have H2 running as an embedded database for smaller e-commerce clients
> in production without issues.
>
the main point for me is, for which solution we have the most experience
(wo)man power to implement it.
Kind regards
> Best regards,
> Carl
>
>>
>>> On Dec 7, 2020, at 2:46 PM, Jörg Schmidt <jo...@j-m-schmidt.de> wrote:
>>>
>>> Hello,
>>>
>>>> -----Original Message-----
>>>> From: Peter Kovacs [mailto:petko@apache.org]
>>>> Sent: Monday, December 07, 2020 7:38 PM
>>>> To: dev@openoffice.apache.org
>>>> Subject: [discussion] future embedded DB in AOO
>>>>
>>>> Hi all,
>>>>
>>>> I would like to know how you guys feel we should move on with
>>>> our base
>>>> Database. ...
>>> There is no compelling reason for us to join LO on the database
>>> issue, it may even be strategically smart not to.
>>>
>>> I know neither H2 nor Apache Derby and have only briefly checked
>>> Wikipedia and if these two were available I would be in favour of
>>> Derby, because:
>>> -obviously better support for SQL
>>> -maybe, due to the longer history, the more mature system
>>> -apparently better encryption features
>>> -Apache Project
>>>
>>> But maybe there is an argument for H2:
>>> this project seems to be, in essence, driven by a single developer,
>>> most likely the one who would welcome H2 to become more known when we
>>> use it in AOO and would be willing to coordinate the development of
>>> H2 with us, so that H2 adapts to the requirements of AOO in the best
>>> possible way.
>>>
>>>
>>> With both databases, however, Java bothers me - or have we given up
>>> the intention of making AOO less dependent on Java?
>>> (yes, of course this is a very long-term question)
>>>
>>>
>>>
>>>
>>> greetings,
>>> Jörg
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
>>> <ma...@openoffice.apache.org>
>>> For additional commands, e-mail: dev-help@openoffice.apache.org
>>> <ma...@openoffice.apache.org>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
> For additional commands, e-mail: dev-help@openoffice.apache.org
>
--
Mechtilde Stehmann
## Apache OpenOffice
## Freie Office Suite für Linux, MacOSX, Windows und OS/2
## Debian Developer
## PGP encryption welcome
## F0E3 7F3D C87A 4998 2899 39E7 F287 7BBA 141A AD7F
Re: [discussion] future embedded DB in AOO
Posted by Carl Marcum <cm...@apache.org>.
Hi All,
On 12/7/20 4:11 PM, Jim Jagielski wrote:
> I agree that eating our own dogfood and using Derby would be the "better" path to take; plus it is relatively small while at the same time being quite powerful and compliant.
It's not well known but Derby is also rebranded and included in the JDK
as JavaDB.
I have used Derby in Embedded and also Client/Server modes and found it
to be a good solution.
I do a lot of Grails development and Grails uses H2 in development mode
and it has a nice web console available.
I have H2 running as an embedded database for smaller e-commerce clients
in production without issues.
Best regards,
Carl
>
>> On Dec 7, 2020, at 2:46 PM, Jörg Schmidt <jo...@j-m-schmidt.de> wrote:
>>
>> Hello,
>>
>>> -----Original Message-----
>>> From: Peter Kovacs [mailto:petko@apache.org]
>>> Sent: Monday, December 07, 2020 7:38 PM
>>> To: dev@openoffice.apache.org
>>> Subject: [discussion] future embedded DB in AOO
>>>
>>> Hi all,
>>>
>>> I would like to know how you guys feel we should move on with
>>> our base
>>> Database. ...
>> There is no compelling reason for us to join LO on the database issue, it may even be strategically smart not to.
>>
>> I know neither H2 nor Apache Derby and have only briefly checked Wikipedia and if these two were available I would be in favour of Derby, because:
>> -obviously better support for SQL
>> -maybe, due to the longer history, the more mature system
>> -apparently better encryption features
>> -Apache Project
>>
>> But maybe there is an argument for H2:
>> this project seems to be, in essence, driven by a single developer, most likely the one who would welcome H2 to become more known when we use it in AOO and would be willing to coordinate the development of H2 with us, so that H2 adapts to the requirements of AOO in the best possible way.
>>
>>
>> With both databases, however, Java bothers me - or have we given up the intention of making AOO less dependent on Java?
>> (yes, of course this is a very long-term question)
>>
>>
>>
>>
>> greetings,
>> Jörg
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org <ma...@openoffice.apache.org>
>> For additional commands, e-mail: dev-help@openoffice.apache.org <ma...@openoffice.apache.org>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: [discussion] future embedded DB in AOO
Posted by Jim Jagielski <ji...@jaguNET.com>.
I agree that eating our own dogfood and using Derby would be the "better" path to take; plus it is relatively small while at the same time being quite powerful and compliant.
> On Dec 7, 2020, at 2:46 PM, Jörg Schmidt <jo...@j-m-schmidt.de> wrote:
>
> Hello,
>
>> -----Original Message-----
>> From: Peter Kovacs [mailto:petko@apache.org]
>> Sent: Monday, December 07, 2020 7:38 PM
>> To: dev@openoffice.apache.org
>> Subject: [discussion] future embedded DB in AOO
>>
>> Hi all,
>>
>> I would like to know how you guys feel we should move on with
>> our base
>> Database. ...
>
> There is no compelling reason for us to join LO on the database issue, it may even be strategically smart not to.
>
> I know neither H2 nor Apache Derby and have only briefly checked Wikipedia and if these two were available I would be in favour of Derby, because:
> -obviously better support for SQL
> -maybe, due to the longer history, the more mature system
> -apparently better encryption features
> -Apache Project
>
> But maybe there is an argument for H2:
> this project seems to be, in essence, driven by a single developer, most likely the one who would welcome H2 to become more known when we use it in AOO and would be willing to coordinate the development of H2 with us, so that H2 adapts to the requirements of AOO in the best possible way.
>
>
> With both databases, however, Java bothers me - or have we given up the intention of making AOO less dependent on Java?
> (yes, of course this is a very long-term question)
>
>
>
>
> greetings,
> Jörg
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org <ma...@openoffice.apache.org>
> For additional commands, e-mail: dev-help@openoffice.apache.org <ma...@openoffice.apache.org>
Re: [discussion] future embedded DB in AOO
Posted by Jim Jagielski <ji...@jaguNET.com>.
> On Dec 7, 2020, at 2:46 PM, Jörg Schmidt <jo...@j-m-schmidt.de> wrote:
>
> With both databases, however, Java bothers me - or have we given up the intention of making AOO less dependent on Java?
> (yes, of course this is a very long-term question)
But still a good one... personally, SQlite seems like a better, more strategic, option with that in mind.
RE: [discussion] future embedded DB in AOO
Posted by Jörg Schmidt <jo...@j-m-schmidt.de>.
Hello,
> -----Original Message-----
> From: Peter Kovacs [mailto:petko@apache.org]
> Sent: Monday, December 07, 2020 7:38 PM
> To: dev@openoffice.apache.org
> Subject: [discussion] future embedded DB in AOO
>
> Hi all,
>
> I would like to know how you guys feel we should move on with
> our base
> Database. ...
There is no compelling reason for us to join LO on the database issue, it may even be strategically smart not to.
I know neither H2 nor Apache Derby and have only briefly checked Wikipedia and if these two were available I would be in favour of Derby, because:
-obviously better support for SQL
-maybe, due to the longer history, the more mature system
-apparently better encryption features
-Apache Project
But maybe there is an argument for H2:
this project seems to be, in essence, driven by a single developer, most likely the one who would welcome H2 to become more known when we use it in AOO and would be willing to coordinate the development of H2 with us, so that H2 adapts to the requirements of AOO in the best possible way.
With both databases, however, Java bothers me - or have we given up the intention of making AOO less dependent on Java?
(yes, of course this is a very long-term question)
greetings,
Jörg
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
RE: [discussion] future embedded DB in AOO
Posted by Jörg Schmidt <jo...@j-m-schmidt.de>.
Hello,
> -----Original Message-----
> From: Mechtilde [mailto:ooo@mechtilde.de]
> Sent: Monday, December 07, 2020 8:11 PM
> To: dev@openoffice.apache.org
> Subject: Re: [discussion] future embedded DB in AOO
>
> Hello Peter,
>
> Am 07.12.20 um 19:38 schrieb Peter Kovacs:
> > Hi all,
> >
> > I would like to know how you guys feel we should move on
> with our base
> > Database. Our current strategy is a small sized embedded DB
> using HSQLDB
> > (BSD). There are 2 other options in the same weight class,
> that is H2
> > (EPL1 or MPL2) and Apache Derby (AL2).
>
> I prefer H2by several reasens:
> 1. H2 is the continued deelopment of HSQLDB
> 2. H2 is also used in other applications like Hibiscus (and
> JVerein ;-))
> 3. I know also some proprietary software which use it as an
> embedded DB
If I read wikipedia the SQL support seems to be worse (in the sense of incomplete) for H2 than for Derby.
Is there anything more to say about this? (Maybe the impression I get from wikipedia is deceptive?)
greetings,
Jörg
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: [discussion] future embedded DB in AOO
Posted by Mechtilde <oo...@mechtilde.de>.
Hello Peter,
Am 07.12.20 um 19:38 schrieb Peter Kovacs:
> Hi all,
>
> I would like to know how you guys feel we should move on with our base
> Database. Our current strategy is a small sized embedded DB using HSQLDB
> (BSD). There are 2 other options in the same weight class, that is H2
> (EPL1 or MPL2) and Apache Derby (AL2).
I prefer H2by several reasens:
1. H2 is the continued deelopment of HSQLDB
2. H2 is also used in other applications like Hibiscus (and JVerein ;-))
3. I know also some proprietary software which use it as an embedded DB
kind regards
Mechtilde
>
> Something like SQLLite (PublicDomain) could be technical interesting,
> but I think their licensing is not so appealing.
>
> We could also consider firebird (MPL) like LO, but I have the impression
> that this is not a lightweight DB, but featurerich. And I wonder if this
> is really something we should provide. I would rather develop a way to
> make it easy to switch the DB when a Project grows.
>
> What are your thoughts?
>
>
> All the Best
>
> Peter
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
> For additional commands, e-mail: dev-help@openoffice.apache.org
>
--
Mechtilde Stehmann
## Apache OpenOffice
## Freie Office Suite für Linux, MacOSX, Windows und OS/2
## Debian Developer
## PGP encryption welcome
## F0E3 7F3D C87A 4998 2899 39E7 F287 7BBA 141A AD7F
Re: Java, OOo, ... (Re: [discussion] future embedded DB in AOO
Posted by Rony G Flatscher <ro...@wu.ac.at>.
Thank you as well, also for the link which I was not aware of! 😊
—-rony
Rony G. Flatscher (mobil/e)
> Am 10.12.2020 um 12:49 schrieb Matthias Seidel <ma...@hamburg.de>:
>
> Hi all,
>
>> Am 10.12.20 um 01:59 schrieb Peter Kovacs:
>> Thank you Rony for the Java deep dive.
>
> +1
>
> And BTW:
>
> https://www.youtube.com/watch?v=p4DeOilMOQA
>
> Regards,
>
> Matthias
>
>>
>>> On 08.12.20 19:11, Rony G. Flatscher (Apache) wrote:
>>> Hmm, just some thoughts and pointers that may be helpful in the
>>> context of the current Java and
>>> possibly HSQL discussions.
>>>
>>> Ad Java and OpenOffice
>>>
>>> IMHO Java has been fulfilling Sun's original promise "write
>>> once, run everywhere" in a very
>>> impressive way for decades by now! When Sun (being the inventor
>>> of Java) bought Star Division
>>> to acquire StarOffice and making the source also available via
>>> openoffice.org, Sun also had the
>>> Java interfaces put into the office suite, such that Java
>>> programmers would be able to interface
>>> with OOo (using "OOo" to also mean AOO and the LO-fork) via UNO
>>> to this very day, and also
>>> enabling Java to become an additional programming language to
>>> C++ to create OOo packages. This
>>> has been some achievement and to me has been impressive to this
>>> very day!
>>>
>>> Java AOO components and Java applications interacting with AOO
>>> have been working for almost
>>> decades without a need of change! (Also such applications would
>>> be deployable on different
>>> operating systems without a need to rewrite and recompile them!)
>>>
>>> Discussing removing Java dependencies given this history and
>>> integration makes me a little bit
>>> nervous if looking at some of the arguments, which may go back
>>> to misconceptions, hear-say and
>>> possibly wrong information published about Java (like Oracle's
>>> change in its license would make
>>> it impossible for software XXX to continue to use it ... which
>>> simply is not the case).
>>>
>>> The scripting interface to OOo is written in Java. Therefore it
>>> has become possible to use the
>>> Java implemented scripting languages JavaScript and BeanShell to
>>> create OOo macros, besides OOo
>>> Basic and Python. In addition - as others have done also - I
>>> authored a little package that uses
>>> the OOo scripting interface in order to make another scripting
>>> language (ooRexx [1]) available
>>> as an OOo macro language (one package creates a bidirectional
>>> bridge between ooRexx and Java
>>> [2], the other package camouflages UNO as ooRexx and exploits
>>> the UNO reflection mechanism to
>>> make it easy on programmers to interface with OOo [3]), so this
>>> is one reason, why being a
>>> little bit nervous about a possible wrong assessment of Java...
>>>
>>> Ad Java Distributions
>>>
>>> It seems that many people think that Java is only available in a
>>> proprietary form from Oracle
>>> [4], which changed its license terms for its Java 1.8/8 and up,
>>> around the time when modular
>>> Java (Java 9) got released. Oracle is regarded to be the owner
>>> of Java and therefore people tend
>>> to think that there is no open-source, free alternative
>>> available, which is not correct.
>>>
>>> Enter "OpenJDK" [5] the open-source version of Java: this allows
>>> you to use Java for free (it is
>>> GPL with the CLASSPATH exception license such that your code
>>> using it does not fall under GPL
>>> automatically). With the modular versions of Java, starting with
>>> Java 9, one is even able to
>>> create one own's Java (runtime environment, JRE) by combining
>>> the Java modules one wishes to
>>> deploy[6]. (This would even open up the opportunity for OOo to
>>> create and distribute its own
>>> tailored JRE, should such a need arise.)
>>>
>>> If you do not feel inclined to create your own JRE (Java
>>> runtime-environment) then you can
>>> download Java=OpenJDK JRE for your particular platform from e.g.
>>> AdoptOpenJDK [7], Liberica [8]
>>> or Zulu [9] to name a few.
>>>
>>> So Java is available in an open-source and free form with the
>>> term OpenJDK (open Java
>>> development kit) [5].
>>>
>>> One tidbit in this "license" context: programmers who wish to
>>> contribute to Java/OpenJDK must
>>> sign an "OCA" (Oracle contributer agreement [15]) giving more or
>>> less all rights on the
>>> contributed software to Oracle which then makes the software
>>> available to the OpenJDK with GPL
>>> and the CLASSPATH exception. Something that Sun had done with
>>> software contributed to OOo. This
>>> BTW allowed later Oracle (after buying Sun and acquiring all of
>>> Sun's software rights) to
>>> contribute the OOo source code to the Apache software
>>> foundation, making it in the end possible
>>> to create and release AOO under the Apache license! (BTW, one
>>> would be able to release one
>>> own's, contributed code with additional, different licenses.
>>> Something that would also be
>>> possible for e.g. LO-contributors, but many are not aware of
>>> this it seems.)
>>>
>>> Ad Java 1.8/8 Versus Java 9 and Later ...
>>>
>>> Java 9 got introduced in the fall of 2017 [10]. There are a few
>>> notable changes:
>>>
>>> * one being that Java has been finally modularized: internal
>>> and reflective code now is
>>> access-based (using the package java.lang.invoke), such that
>>> setAccessible as used in the
>>> prior versions via the java.lang.reflect package in order to
>>> invoke reflectively will not be
>>> allowed anymore. In the transition phase this may cause
>>> many, many problems with code that
>>> was created prior to Java 9, such that for some time the
>>> modularized versions of Java will
>>> tolerate access via java.lang.reflect as in the past (and
>>> there are even provisions in place
>>> to have command-line arguments to extend these permissions
>>> for a longer period with newer
>>> versions of modularized Java). (Where java.lang.reflect
>>> would be usable to invoke a method
>>> in the otherwise inaccessilbe sun-package, java.lang.invoke
>>> will not allow that if the
>>> module is not exported, rather one needs to go up the
>>> inheritance tree to find a class in an
>>> accessible/exported module and invoke the method via that
>>> class' method object),
>>>
>>> * one can create one own's Java runtime environment (JRE) from
>>> the Java modules [6];
>>> Java/OpenJDK 11 for instance does not contain the four
>>> JavaFX modules by default (a
>>> separate, active project, OpenJFX [16]) anymore;
>>> Java/OpenJDK 15 removes Nashorn
>>> (JavaScript), however at the end of November a proper
>>> Nashorn (JavaScript) module was made
>>> available that can be mixed into one own's modular JRE
>>> (again look at the sample in [6] to
>>> save yourself some research time) of Java 15 or later,
>>>
>>> * the release cycle of Java/OpenJDK had been changed to six
>>> (!) months: hence there have been
>>> two major releases of Java/OpenJDK per year since then, the
>>> latest one being Java/OpenJDK 15
>>> as of September 2020. One purpose supposedly was to make it
>>> possible to add new features to
>>> the Java/OpenJDK language quicker than in the past. The LTS
>>> (long-term support) versions are
>>> then roughly comparable to the classic release cycles of
>>> Java/OpenJDK, cf. [10].
>>>
>>> If you look at the Java/OpenJDK versions [10] you will see
>>> "LTS", where Java 8 will be serviced
>>> until 2030 (!) and the modular LTS version Java/OpenJDK 11 will
>>> be serviced until September
>>> 2026, another six years from now.
>>>
>>> Ad HSQL
>>>
>>> It seems that the rather old version of AOO's HSQL may have been
>>> caused by sticking to older
>>> versions of Java. HSQL [11] states that the HSQL version 2.3.8
>>> is based on Java 1.6/6, and its
>>> latest HSQL version 2.5.1 (as of June 2020) is based on
>>> Java/OpenJDK 1.8/8. No idea whether
>>> 2.3.8 or 2.5.1 break with the older APIs that the current AOO
>>> employs. (There is an interesting
>>> comparison of the properties of HSQL and H2 [12] at [13]. For
>>> completeness in the context of the
>>> discussions, the homepage of Apache Derby: [14].)
>>>
>>> With this information, hopefully, Java remains as a first citizen in
>>> AOO!
>>>
>>> Ceterum censeo, Java should remain as a first citizen in AOO!
>>> :)
>>>
>>> HTH,
>>>
>>> ---rony
>>>
>>>
>>> [1] ooRexx (implemented in C++):
>>> <https://sourceforge.net/projects/oorexx/files/oorexx/5.0.0beta/>
>>>
>>> [2] BSF4ooRexx (Java bridge a mixture of C++, Java and ooRexx):
>>> <https://sourceforge.net/projects/bsf4oorexx/files/beta/20200928/>
>>>
>>> [3] "Scripting Apache Office", slides, ApacheCon Europe 2012, Germany:
>>> <AutoJava-BSF4ooRexx-04-OOo-ACE12.pdf >, UNO programming, programming
>>> swriter, scalc, sdraw,
>>> simpress, URE (UNO runtime environment)
>>>
>>> [4] Oracle's Java: <https://java.com/>
>>>
>>> [5] OpenJDK: <https://openjdk.java.net/>
>>>
>>> [6] "Setup Environment for Java/OpenJDK (2020-05-29)", slides, WU
>>> Vienna/Austria:
>>> <http://wi.wu.ac.at/rgf/wu/lehre/autojava/material/foils/AutoJava-BSF4ooRexx-08-Environment.pdf
>>>>
>>>
>>> [7] "AdoptOpenJDK" OpenJDK/Java: <https://adoptOpenJDK.net>; this
>>> version even allows you to even
>>> pick the JVM to run Java
>>>
>>> [8] "Liberica" OpenJDK/Java: <https://bell-sw.com/>, look for "full
>>> Liberica", if you want the
>>> JavaFX modules to be available
>>>
>>> [9] "Zulu" OpenJDK/Java: <https://www.azul.com/downloads/>, look for
>>> "JDK FX", if you want the
>>> JavaFX modules to be available
>>>
>>> [10] "Java version history", Wikipedia:
>>> <https://en.wikipedia.org/wiki/Java_version_history
>>>
>>> [11] "HyperSQL", Homepage: <http://hsqldb.org/>
>>>
>>> [12] "H2 Database Engine", Homepage: <https://www.h2database.com/>
>>>
>>> [13] "System Properties Comparison H2 vs. HyperSQL", a current article:
>>> <https://db-engines.com/en/system/H2%3BHyperSQL>
>>>
>>> [14] "Apache Derby", homepage: <https://db.apache.org/derby/>
>>>
>>> [15] "Oracle's Contributor Agreement (OCA)":
>>> <https://www.oracle.com/technical-resources/oracle-contributor-agreement.html>
>>>
>>>
>>> [16] "OpenJFX", homepage: <https://openjfx.io/
>>>
>>>
>>>
>>>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: Java, OOo, ... (Re: [discussion] future embedded DB in AOO
Posted by Matthias Seidel <ma...@hamburg.de>.
Hi all,
Am 10.12.20 um 01:59 schrieb Peter Kovacs:
> Thank you Rony for the Java deep dive.
+1
And BTW:
https://www.youtube.com/watch?v=p4DeOilMOQA
Regards,
Matthias
>
> On 08.12.20 19:11, Rony G. Flatscher (Apache) wrote:
>> Hmm, just some thoughts and pointers that may be helpful in the
>> context of the current Java and
>> possibly HSQL discussions.
>>
>> Ad Java and OpenOffice
>>
>> IMHO Java has been fulfilling Sun's original promise "write
>> once, run everywhere" in a very
>> impressive way for decades by now! When Sun (being the inventor
>> of Java) bought Star Division
>> to acquire StarOffice and making the source also available via
>> openoffice.org, Sun also had the
>> Java interfaces put into the office suite, such that Java
>> programmers would be able to interface
>> with OOo (using "OOo" to also mean AOO and the LO-fork) via UNO
>> to this very day, and also
>> enabling Java to become an additional programming language to
>> C++ to create OOo packages. This
>> has been some achievement and to me has been impressive to this
>> very day!
>>
>> Java AOO components and Java applications interacting with AOO
>> have been working for almost
>> decades without a need of change! (Also such applications would
>> be deployable on different
>> operating systems without a need to rewrite and recompile them!)
>>
>> Discussing removing Java dependencies given this history and
>> integration makes me a little bit
>> nervous if looking at some of the arguments, which may go back
>> to misconceptions, hear-say and
>> possibly wrong information published about Java (like Oracle's
>> change in its license would make
>> it impossible for software XXX to continue to use it ... which
>> simply is not the case).
>>
>> The scripting interface to OOo is written in Java. Therefore it
>> has become possible to use the
>> Java implemented scripting languages JavaScript and BeanShell to
>> create OOo macros, besides OOo
>> Basic and Python. In addition - as others have done also - I
>> authored a little package that uses
>> the OOo scripting interface in order to make another scripting
>> language (ooRexx [1]) available
>> as an OOo macro language (one package creates a bidirectional
>> bridge between ooRexx and Java
>> [2], the other package camouflages UNO as ooRexx and exploits
>> the UNO reflection mechanism to
>> make it easy on programmers to interface with OOo [3]), so this
>> is one reason, why being a
>> little bit nervous about a possible wrong assessment of Java...
>>
>> Ad Java Distributions
>>
>> It seems that many people think that Java is only available in a
>> proprietary form from Oracle
>> [4], which changed its license terms for its Java 1.8/8 and up,
>> around the time when modular
>> Java (Java 9) got released. Oracle is regarded to be the owner
>> of Java and therefore people tend
>> to think that there is no open-source, free alternative
>> available, which is not correct.
>>
>> Enter "OpenJDK" [5] the open-source version of Java: this allows
>> you to use Java for free (it is
>> GPL with the CLASSPATH exception license such that your code
>> using it does not fall under GPL
>> automatically). With the modular versions of Java, starting with
>> Java 9, one is even able to
>> create one own's Java (runtime environment, JRE) by combining
>> the Java modules one wishes to
>> deploy[6]. (This would even open up the opportunity for OOo to
>> create and distribute its own
>> tailored JRE, should such a need arise.)
>>
>> If you do not feel inclined to create your own JRE (Java
>> runtime-environment) then you can
>> download Java=OpenJDK JRE for your particular platform from e.g.
>> AdoptOpenJDK [7], Liberica [8]
>> or Zulu [9] to name a few.
>>
>> So Java is available in an open-source and free form with the
>> term OpenJDK (open Java
>> development kit) [5].
>>
>> One tidbit in this "license" context: programmers who wish to
>> contribute to Java/OpenJDK must
>> sign an "OCA" (Oracle contributer agreement [15]) giving more or
>> less all rights on the
>> contributed software to Oracle which then makes the software
>> available to the OpenJDK with GPL
>> and the CLASSPATH exception. Something that Sun had done with
>> software contributed to OOo. This
>> BTW allowed later Oracle (after buying Sun and acquiring all of
>> Sun's software rights) to
>> contribute the OOo source code to the Apache software
>> foundation, making it in the end possible
>> to create and release AOO under the Apache license! (BTW, one
>> would be able to release one
>> own's, contributed code with additional, different licenses.
>> Something that would also be
>> possible for e.g. LO-contributors, but many are not aware of
>> this it seems.)
>>
>> Ad Java 1.8/8 Versus Java 9 and Later ...
>>
>> Java 9 got introduced in the fall of 2017 [10]. There are a few
>> notable changes:
>>
>> * one being that Java has been finally modularized: internal
>> and reflective code now is
>> access-based (using the package java.lang.invoke), such that
>> setAccessible as used in the
>> prior versions via the java.lang.reflect package in order to
>> invoke reflectively will not be
>> allowed anymore. In the transition phase this may cause
>> many, many problems with code that
>> was created prior to Java 9, such that for some time the
>> modularized versions of Java will
>> tolerate access via java.lang.reflect as in the past (and
>> there are even provisions in place
>> to have command-line arguments to extend these permissions
>> for a longer period with newer
>> versions of modularized Java). (Where java.lang.reflect
>> would be usable to invoke a method
>> in the otherwise inaccessilbe sun-package, java.lang.invoke
>> will not allow that if the
>> module is not exported, rather one needs to go up the
>> inheritance tree to find a class in an
>> accessible/exported module and invoke the method via that
>> class' method object),
>>
>> * one can create one own's Java runtime environment (JRE) from
>> the Java modules [6];
>> Java/OpenJDK 11 for instance does not contain the four
>> JavaFX modules by default (a
>> separate, active project, OpenJFX [16]) anymore;
>> Java/OpenJDK 15 removes Nashorn
>> (JavaScript), however at the end of November a proper
>> Nashorn (JavaScript) module was made
>> available that can be mixed into one own's modular JRE
>> (again look at the sample in [6] to
>> save yourself some research time) of Java 15 or later,
>>
>> * the release cycle of Java/OpenJDK had been changed to six
>> (!) months: hence there have been
>> two major releases of Java/OpenJDK per year since then, the
>> latest one being Java/OpenJDK 15
>> as of September 2020. One purpose supposedly was to make it
>> possible to add new features to
>> the Java/OpenJDK language quicker than in the past. The LTS
>> (long-term support) versions are
>> then roughly comparable to the classic release cycles of
>> Java/OpenJDK, cf. [10].
>>
>> If you look at the Java/OpenJDK versions [10] you will see
>> "LTS", where Java 8 will be serviced
>> until 2030 (!) and the modular LTS version Java/OpenJDK 11 will
>> be serviced until September
>> 2026, another six years from now.
>>
>> Ad HSQL
>>
>> It seems that the rather old version of AOO's HSQL may have been
>> caused by sticking to older
>> versions of Java. HSQL [11] states that the HSQL version 2.3.8
>> is based on Java 1.6/6, and its
>> latest HSQL version 2.5.1 (as of June 2020) is based on
>> Java/OpenJDK 1.8/8. No idea whether
>> 2.3.8 or 2.5.1 break with the older APIs that the current AOO
>> employs. (There is an interesting
>> comparison of the properties of HSQL and H2 [12] at [13]. For
>> completeness in the context of the
>> discussions, the homepage of Apache Derby: [14].)
>>
>> With this information, hopefully, Java remains as a first citizen in
>> AOO!
>>
>> Ceterum censeo, Java should remain as a first citizen in AOO!
>> :)
>>
>> HTH,
>>
>> ---rony
>>
>>
>> [1] ooRexx (implemented in C++):
>> <https://sourceforge.net/projects/oorexx/files/oorexx/5.0.0beta/>
>>
>> [2] BSF4ooRexx (Java bridge a mixture of C++, Java and ooRexx):
>> <https://sourceforge.net/projects/bsf4oorexx/files/beta/20200928/>
>>
>> [3] "Scripting Apache Office", slides, ApacheCon Europe 2012, Germany:
>> <AutoJava-BSF4ooRexx-04-OOo-ACE12.pdf >, UNO programming, programming
>> swriter, scalc, sdraw,
>> simpress, URE (UNO runtime environment)
>>
>> [4] Oracle's Java: <https://java.com/>
>>
>> [5] OpenJDK: <https://openjdk.java.net/>
>>
>> [6] "Setup Environment for Java/OpenJDK (2020-05-29)", slides, WU
>> Vienna/Austria:
>> <http://wi.wu.ac.at/rgf/wu/lehre/autojava/material/foils/AutoJava-BSF4ooRexx-08-Environment.pdf
>> >
>>
>> [7] "AdoptOpenJDK" OpenJDK/Java: <https://adoptOpenJDK.net>; this
>> version even allows you to even
>> pick the JVM to run Java
>>
>> [8] "Liberica" OpenJDK/Java: <https://bell-sw.com/>, look for "full
>> Liberica", if you want the
>> JavaFX modules to be available
>>
>> [9] "Zulu" OpenJDK/Java: <https://www.azul.com/downloads/>, look for
>> "JDK FX", if you want the
>> JavaFX modules to be available
>>
>> [10] "Java version history", Wikipedia:
>> <https://en.wikipedia.org/wiki/Java_version_history
>>
>> [11] "HyperSQL", Homepage: <http://hsqldb.org/>
>>
>> [12] "H2 Database Engine", Homepage: <https://www.h2database.com/>
>>
>> [13] "System Properties Comparison H2 vs. HyperSQL", a current article:
>> <https://db-engines.com/en/system/H2%3BHyperSQL>
>>
>> [14] "Apache Derby", homepage: <https://db.apache.org/derby/>
>>
>> [15] "Oracle's Contributor Agreement (OCA)":
>> <https://www.oracle.com/technical-resources/oracle-contributor-agreement.html>
>>
>>
>> [16] "OpenJFX", homepage: <https://openjfx.io/
>>
>>
>>
>>
Re: Java, OOo, ... (Re: [discussion] future embedded DB in AOO
Posted by Peter Kovacs <pe...@apache.org>.
Thank you Rony for the Java deep dive.
On 08.12.20 19:11, Rony G. Flatscher (Apache) wrote:
> Hmm, just some thoughts and pointers that may be helpful in the context of the current Java and
> possibly HSQL discussions.
>
> Ad Java and OpenOffice
>
> IMHO Java has been fulfilling Sun's original promise "write once, run everywhere" in a very
> impressive way for decades by now! When Sun (being the inventor of Java) bought Star Division
> to acquire StarOffice and making the source also available via openoffice.org, Sun also had the
> Java interfaces put into the office suite, such that Java programmers would be able to interface
> with OOo (using "OOo" to also mean AOO and the LO-fork) via UNO to this very day, and also
> enabling Java to become an additional programming language to C++ to create OOo packages. This
> has been some achievement and to me has been impressive to this very day!
>
> Java AOO components and Java applications interacting with AOO have been working for almost
> decades without a need of change! (Also such applications would be deployable on different
> operating systems without a need to rewrite and recompile them!)
>
> Discussing removing Java dependencies given this history and integration makes me a little bit
> nervous if looking at some of the arguments, which may go back to misconceptions, hear-say and
> possibly wrong information published about Java (like Oracle's change in its license would make
> it impossible for software XXX to continue to use it ... which simply is not the case).
>
> The scripting interface to OOo is written in Java. Therefore it has become possible to use the
> Java implemented scripting languages JavaScript and BeanShell to create OOo macros, besides OOo
> Basic and Python. In addition - as others have done also - I authored a little package that uses
> the OOo scripting interface in order to make another scripting language (ooRexx [1]) available
> as an OOo macro language (one package creates a bidirectional bridge between ooRexx and Java
> [2], the other package camouflages UNO as ooRexx and exploits the UNO reflection mechanism to
> make it easy on programmers to interface with OOo [3]), so this is one reason, why being a
> little bit nervous about a possible wrong assessment of Java...
>
> Ad Java Distributions
>
> It seems that many people think that Java is only available in a proprietary form from Oracle
> [4], which changed its license terms for its Java 1.8/8 and up, around the time when modular
> Java (Java 9) got released. Oracle is regarded to be the owner of Java and therefore people tend
> to think that there is no open-source, free alternative available, which is not correct.
>
> Enter "OpenJDK" [5] the open-source version of Java: this allows you to use Java for free (it is
> GPL with the CLASSPATH exception license such that your code using it does not fall under GPL
> automatically). With the modular versions of Java, starting with Java 9, one is even able to
> create one own's Java (runtime environment, JRE) by combining the Java modules one wishes to
> deploy[6]. (This would even open up the opportunity for OOo to create and distribute its own
> tailored JRE, should such a need arise.)
>
> If you do not feel inclined to create your own JRE (Java runtime-environment) then you can
> download Java=OpenJDK JRE for your particular platform from e.g. AdoptOpenJDK [7], Liberica [8]
> or Zulu [9] to name a few.
>
> So Java is available in an open-source and free form with the term OpenJDK (open Java
> development kit) [5].
>
> One tidbit in this "license" context: programmers who wish to contribute to Java/OpenJDK must
> sign an "OCA" (Oracle contributer agreement [15]) giving more or less all rights on the
> contributed software to Oracle which then makes the software available to the OpenJDK with GPL
> and the CLASSPATH exception. Something that Sun had done with software contributed to OOo. This
> BTW allowed later Oracle (after buying Sun and acquiring all of Sun's software rights) to
> contribute the OOo source code to the Apache software foundation, making it in the end possible
> to create and release AOO under the Apache license! (BTW, one would be able to release one
> own's, contributed code with additional, different licenses. Something that would also be
> possible for e.g. LO-contributors, but many are not aware of this it seems.)
>
> Ad Java 1.8/8 Versus Java 9 and Later ...
>
> Java 9 got introduced in the fall of 2017 [10]. There are a few notable changes:
>
> * one being that Java has been finally modularized: internal and reflective code now is
> access-based (using the package java.lang.invoke), such that setAccessible as used in the
> prior versions via the java.lang.reflect package in order to invoke reflectively will not be
> allowed anymore. In the transition phase this may cause many, many problems with code that
> was created prior to Java 9, such that for some time the modularized versions of Java will
> tolerate access via java.lang.reflect as in the past (and there are even provisions in place
> to have command-line arguments to extend these permissions for a longer period with newer
> versions of modularized Java). (Where java.lang.reflect would be usable to invoke a method
> in the otherwise inaccessilbe sun-package, java.lang.invoke will not allow that if the
> module is not exported, rather one needs to go up the inheritance tree to find a class in an
> accessible/exported module and invoke the method via that class' method object),
>
> * one can create one own's Java runtime environment (JRE) from the Java modules [6];
> Java/OpenJDK 11 for instance does not contain the four JavaFX modules by default (a
> separate, active project, OpenJFX [16]) anymore; Java/OpenJDK 15 removes Nashorn
> (JavaScript), however at the end of November a proper Nashorn (JavaScript) module was made
> available that can be mixed into one own's modular JRE (again look at the sample in [6] to
> save yourself some research time) of Java 15 or later,
>
> * the release cycle of Java/OpenJDK had been changed to six (!) months: hence there have been
> two major releases of Java/OpenJDK per year since then, the latest one being Java/OpenJDK 15
> as of September 2020. One purpose supposedly was to make it possible to add new features to
> the Java/OpenJDK language quicker than in the past. The LTS (long-term support) versions are
> then roughly comparable to the classic release cycles of Java/OpenJDK, cf. [10].
>
> If you look at the Java/OpenJDK versions [10] you will see "LTS", where Java 8 will be serviced
> until 2030 (!) and the modular LTS version Java/OpenJDK 11 will be serviced until September
> 2026, another six years from now.
>
> Ad HSQL
>
> It seems that the rather old version of AOO's HSQL may have been caused by sticking to older
> versions of Java. HSQL [11] states that the HSQL version 2.3.8 is based on Java 1.6/6, and its
> latest HSQL version 2.5.1 (as of June 2020) is based on Java/OpenJDK 1.8/8. No idea whether
> 2.3.8 or 2.5.1 break with the older APIs that the current AOO employs. (There is an interesting
> comparison of the properties of HSQL and H2 [12] at [13]. For completeness in the context of the
> discussions, the homepage of Apache Derby: [14].)
>
> With this information, hopefully, Java remains as a first citizen in AOO!
>
> Ceterum censeo, Java should remain as a first citizen in AOO!
> :)
>
> HTH,
>
> ---rony
>
>
> [1] ooRexx (implemented in C++): <https://sourceforge.net/projects/oorexx/files/oorexx/5.0.0beta/>
>
> [2] BSF4ooRexx (Java bridge a mixture of C++, Java and ooRexx):
> <https://sourceforge.net/projects/bsf4oorexx/files/beta/20200928/>
>
> [3] "Scripting Apache Office", slides, ApacheCon Europe 2012, Germany:
> <AutoJava-BSF4ooRexx-04-OOo-ACE12.pdf >, UNO programming, programming swriter, scalc, sdraw,
> simpress, URE (UNO runtime environment)
>
> [4] Oracle's Java: <https://java.com/>
>
> [5] OpenJDK: <https://openjdk.java.net/>
>
> [6] "Setup Environment for Java/OpenJDK (2020-05-29)", slides, WU Vienna/Austria:
> <http://wi.wu.ac.at/rgf/wu/lehre/autojava/material/foils/AutoJava-BSF4ooRexx-08-Environment.pdf >
>
> [7] "AdoptOpenJDK" OpenJDK/Java: <https://adoptOpenJDK.net>; this version even allows you to even
> pick the JVM to run Java
>
> [8] "Liberica" OpenJDK/Java: <https://bell-sw.com/>, look for "full Liberica", if you want the
> JavaFX modules to be available
>
> [9] "Zulu" OpenJDK/Java: <https://www.azul.com/downloads/>, look for "JDK FX", if you want the
> JavaFX modules to be available
>
> [10] "Java version history", Wikipedia: <https://en.wikipedia.org/wiki/Java_version_history
>
> [11] "HyperSQL", Homepage: <http://hsqldb.org/>
>
> [12] "H2 Database Engine", Homepage: <https://www.h2database.com/>
>
> [13] "System Properties Comparison H2 vs. HyperSQL", a current article:
> <https://db-engines.com/en/system/H2%3BHyperSQL>
>
> [14] "Apache Derby", homepage: <https://db.apache.org/derby/>
>
> [15] "Oracle's Contributor Agreement (OCA)":
> <https://www.oracle.com/technical-resources/oracle-contributor-agreement.html>
>
> [16] "OpenJFX", homepage: <https://openjfx.io/
>
>
>
>
--
This is the Way! http://www.apache.org/theapacheway/index.html
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Java, OOo, ... (Re: [discussion] future embedded DB in AOO
Posted by "Rony G. Flatscher (Apache)" <ro...@apache.org>.
Hmm, just some thoughts and pointers that may be helpful in the context of the current Java and
possibly HSQL discussions.
Ad Java and OpenOffice
IMHO Java has been fulfilling Sun's original promise "write once, run everywhere" in a very
impressive way for decades by now! When Sun (being the inventor of Java) bought Star Division
to acquire StarOffice and making the source also available via openoffice.org, Sun also had the
Java interfaces put into the office suite, such that Java programmers would be able to interface
with OOo (using "OOo" to also mean AOO and the LO-fork) via UNO to this very day, and also
enabling Java to become an additional programming language to C++ to create OOo packages. This
has been some achievement and to me has been impressive to this very day!
Java AOO components and Java applications interacting with AOO have been working for almost
decades without a need of change! (Also such applications would be deployable on different
operating systems without a need to rewrite and recompile them!)
Discussing removing Java dependencies given this history and integration makes me a little bit
nervous if looking at some of the arguments, which may go back to misconceptions, hear-say and
possibly wrong information published about Java (like Oracle's change in its license would make
it impossible for software XXX to continue to use it ... which simply is not the case).
The scripting interface to OOo is written in Java. Therefore it has become possible to use the
Java implemented scripting languages JavaScript and BeanShell to create OOo macros, besides OOo
Basic and Python. In addition - as others have done also - I authored a little package that uses
the OOo scripting interface in order to make another scripting language (ooRexx [1]) available
as an OOo macro language (one package creates a bidirectional bridge between ooRexx and Java
[2], the other package camouflages UNO as ooRexx and exploits the UNO reflection mechanism to
make it easy on programmers to interface with OOo [3]), so this is one reason, why being a
little bit nervous about a possible wrong assessment of Java...
Ad Java Distributions
It seems that many people think that Java is only available in a proprietary form from Oracle
[4], which changed its license terms for its Java 1.8/8 and up, around the time when modular
Java (Java 9) got released. Oracle is regarded to be the owner of Java and therefore people tend
to think that there is no open-source, free alternative available, which is not correct.
Enter "OpenJDK" [5] the open-source version of Java: this allows you to use Java for free (it is
GPL with the CLASSPATH exception license such that your code using it does not fall under GPL
automatically). With the modular versions of Java, starting with Java 9, one is even able to
create one own's Java (runtime environment, JRE) by combining the Java modules one wishes to
deploy[6]. (This would even open up the opportunity for OOo to create and distribute its own
tailored JRE, should such a need arise.)
If you do not feel inclined to create your own JRE (Java runtime-environment) then you can
download Java=OpenJDK JRE for your particular platform from e.g. AdoptOpenJDK [7], Liberica [8]
or Zulu [9] to name a few.
So Java is available in an open-source and free form with the term OpenJDK (open Java
development kit) [5].
One tidbit in this "license" context: programmers who wish to contribute to Java/OpenJDK must
sign an "OCA" (Oracle contributer agreement [15]) giving more or less all rights on the
contributed software to Oracle which then makes the software available to the OpenJDK with GPL
and the CLASSPATH exception. Something that Sun had done with software contributed to OOo. This
BTW allowed later Oracle (after buying Sun and acquiring all of Sun's software rights) to
contribute the OOo source code to the Apache software foundation, making it in the end possible
to create and release AOO under the Apache license! (BTW, one would be able to release one
own's, contributed code with additional, different licenses. Something that would also be
possible for e.g. LO-contributors, but many are not aware of this it seems.)
Ad Java 1.8/8 Versus Java 9 and Later ...
Java 9 got introduced in the fall of 2017 [10]. There are a few notable changes:
* one being that Java has been finally modularized: internal and reflective code now is
access-based (using the package java.lang.invoke), such that setAccessible as used in the
prior versions via the java.lang.reflect package in order to invoke reflectively will not be
allowed anymore. In the transition phase this may cause many, many problems with code that
was created prior to Java 9, such that for some time the modularized versions of Java will
tolerate access via java.lang.reflect as in the past (and there are even provisions in place
to have command-line arguments to extend these permissions for a longer period with newer
versions of modularized Java). (Where java.lang.reflect would be usable to invoke a method
in the otherwise inaccessilbe sun-package, java.lang.invoke will not allow that if the
module is not exported, rather one needs to go up the inheritance tree to find a class in an
accessible/exported module and invoke the method via that class' method object),
* one can create one own's Java runtime environment (JRE) from the Java modules [6];
Java/OpenJDK 11 for instance does not contain the four JavaFX modules by default (a
separate, active project, OpenJFX [16]) anymore; Java/OpenJDK 15 removes Nashorn
(JavaScript), however at the end of November a proper Nashorn (JavaScript) module was made
available that can be mixed into one own's modular JRE (again look at the sample in [6] to
save yourself some research time) of Java 15 or later,
* the release cycle of Java/OpenJDK had been changed to six (!) months: hence there have been
two major releases of Java/OpenJDK per year since then, the latest one being Java/OpenJDK 15
as of September 2020. One purpose supposedly was to make it possible to add new features to
the Java/OpenJDK language quicker than in the past. The LTS (long-term support) versions are
then roughly comparable to the classic release cycles of Java/OpenJDK, cf. [10].
If you look at the Java/OpenJDK versions [10] you will see "LTS", where Java 8 will be serviced
until 2030 (!) and the modular LTS version Java/OpenJDK 11 will be serviced until September
2026, another six years from now.
Ad HSQL
It seems that the rather old version of AOO's HSQL may have been caused by sticking to older
versions of Java. HSQL [11] states that the HSQL version 2.3.8 is based on Java 1.6/6, and its
latest HSQL version 2.5.1 (as of June 2020) is based on Java/OpenJDK 1.8/8. No idea whether
2.3.8 or 2.5.1 break with the older APIs that the current AOO employs. (There is an interesting
comparison of the properties of HSQL and H2 [12] at [13]. For completeness in the context of the
discussions, the homepage of Apache Derby: [14].)
With this information, hopefully, Java remains as a first citizen in AOO!
Ceterum censeo, Java should remain as a first citizen in AOO!
:)
HTH,
---rony
[1] ooRexx (implemented in C++): <https://sourceforge.net/projects/oorexx/files/oorexx/5.0.0beta/>
[2] BSF4ooRexx (Java bridge a mixture of C++, Java and ooRexx):
<https://sourceforge.net/projects/bsf4oorexx/files/beta/20200928/>
[3] "Scripting Apache Office", slides, ApacheCon Europe 2012, Germany:
<AutoJava-BSF4ooRexx-04-OOo-ACE12.pdf >, UNO programming, programming swriter, scalc, sdraw,
simpress, URE (UNO runtime environment)
[4] Oracle's Java: <https://java.com/>
[5] OpenJDK: <https://openjdk.java.net/>
[6] "Setup Environment for Java/OpenJDK (2020-05-29)", slides, WU Vienna/Austria:
<http://wi.wu.ac.at/rgf/wu/lehre/autojava/material/foils/AutoJava-BSF4ooRexx-08-Environment.pdf >
[7] "AdoptOpenJDK" OpenJDK/Java: <https://adoptOpenJDK.net>; this version even allows you to even
pick the JVM to run Java
[8] "Liberica" OpenJDK/Java: <https://bell-sw.com/>, look for "full Liberica", if you want the
JavaFX modules to be available
[9] "Zulu" OpenJDK/Java: <https://www.azul.com/downloads/>, look for "JDK FX", if you want the
JavaFX modules to be available
[10] "Java version history", Wikipedia: <https://en.wikipedia.org/wiki/Java_version_history
[11] "HyperSQL", Homepage: <http://hsqldb.org/>
[12] "H2 Database Engine", Homepage: <https://www.h2database.com/>
[13] "System Properties Comparison H2 vs. HyperSQL", a current article:
<https://db-engines.com/en/system/H2%3BHyperSQL>
[14] "Apache Derby", homepage: <https://db.apache.org/derby/>
[15] "Oracle's Contributor Agreement (OCA)":
<https://www.oracle.com/technical-resources/oracle-contributor-agreement.html>
[16] "OpenJFX", homepage: <https://openjfx.io/
Re: [discussion] future embedded DB in AOO
Posted by Matthias Seidel <ma...@hamburg.de>.
Hi Damjan,
Am 08.12.20 um 03:06 schrieb Damjan Jovanovic:
> On Mon, Dec 7, 2020 at 8:38 PM Peter Kovacs <pe...@apache.org> wrote:
>
>> Hi all,
>>
>> I would like to know how you guys feel we should move on with our base
>> Database. Our current strategy is a small sized embedded DB using HSQLDB
>> (BSD). There are 2 other options in the same weight class, that is H2
>> (EPL1 or MPL2) and Apache Derby (AL2).
>>
>> Something like SQLLite (PublicDomain) could be technical interesting,
>> but I think their licensing is not so appealing.
>>
>> We could also consider firebird (MPL) like LO, but I have the impression
>> that this is not a lightweight DB, but featurerich. And I wonder if this
>> is really something we should provide. I would rather develop a way to
>> make it easy to switch the DB when a Project grows.
>>
>> What are your thoughts?
>>
>>
> Whatever we do, we have to provide backward compatibility for HSQLDB.
H2 has several compatibility modes icl. HSQLDB:
https://www.h2database.com/html/features.html#feature_list
Regards,
Matthias
>
> I am more interested in something like Apache Calcite, which would let us
> do SQL queries spanning multiple database backends, than in yet another
> database backend.
>
> SQLite is interesting because it's very widely used, runs on iPhone and
> Android, is a single file, performs well, and is lightweight.
>
> The hard part is getting the database to write into .ODB files instead of
> what it natively uses?
>
> Damjan
>
Re: [discussion] future embedded DB in AOO
Posted by Dave Fisher <wa...@apache.org>.
> On Dec 8, 2020, at 2:47 AM, Peter Kovacs <pe...@apache.org> wrote:
>
>
> On 08.12.20 03:06, Damjan Jovanovic wrote:
>> On Mon, Dec 7, 2020 at 8:38 PM Peter Kovacs <pe...@apache.org> wrote:
>>
>>> Hi all,
>>>
>>> I would like to know how you guys feel we should move on with our base
>>> Database. Our current strategy is a small sized embedded DB using HSQLDB
>>> (BSD). There are 2 other options in the same weight class, that is H2
>>> (EPL1 or MPL2) and Apache Derby (AL2).
>>>
>>> Something like SQLLite (PublicDomain) could be technical interesting,
>>> but I think their licensing is not so appealing.
>>>
>>> We could also consider firebird (MPL) like LO, but I have the impression
>>> that this is not a lightweight DB, but featurerich. And I wonder if this
>>> is really something we should provide. I would rather develop a way to
>>> make it easy to switch the DB when a Project grows.
>>>
>>> What are your thoughts?
>>>
>>>
>> Whatever we do, we have to provide backward compatibility for HSQLDB.
>>
>> I am more interested in something like Apache Calcite, which would let us
>> do SQL queries spanning multiple database backends, than in yet another
>> database backend.
>
> This is an interesting approach. However it is not clear to me on the use.
Base can then sit on top of Big Data / NoSQL as an Analytics engine for the rest of us.
>
> We get a half baked DB System and need to add stuff like Data Handling right?
>
> And here we need also to write a loader for ODB files. Would we need an In memory Data representation model, too?
>
> It maybe that we have the basics from Calc, but we still need some work, I think.
>
>>
>> SQLite is interesting because it's very widely used, runs on iPhone and
>> Android, is a single file, performs well, and is lightweight.
>>
>> The hard part is getting the database to write into .ODB files instead of
>> what it natively uses?
> Indeed a topic I did not look into. Actually how is it done today? I do not believe HSSQLDB has a capability for that, right?
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
> For additional commands, e-mail: dev-help@openoffice.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: [discussion] future embedded DB in AOO
Posted by Peter Kovacs <pe...@apache.org>.
On 08.12.20 03:06, Damjan Jovanovic wrote:
> On Mon, Dec 7, 2020 at 8:38 PM Peter Kovacs <pe...@apache.org> wrote:
>
>> Hi all,
>>
>> I would like to know how you guys feel we should move on with our base
>> Database. Our current strategy is a small sized embedded DB using HSQLDB
>> (BSD). There are 2 other options in the same weight class, that is H2
>> (EPL1 or MPL2) and Apache Derby (AL2).
>>
>> Something like SQLLite (PublicDomain) could be technical interesting,
>> but I think their licensing is not so appealing.
>>
>> We could also consider firebird (MPL) like LO, but I have the impression
>> that this is not a lightweight DB, but featurerich. And I wonder if this
>> is really something we should provide. I would rather develop a way to
>> make it easy to switch the DB when a Project grows.
>>
>> What are your thoughts?
>>
>>
> Whatever we do, we have to provide backward compatibility for HSQLDB.
>
> I am more interested in something like Apache Calcite, which would let us
> do SQL queries spanning multiple database backends, than in yet another
> database backend.
This is an interesting approach. However it is not clear to me on the use.
We get a half baked DB System and need to add stuff like Data Handling
right?
And here we need also to write a loader for ODB files. Would we need an
In memory Data representation model, too?
It maybe that we have the basics from Calc, but we still need some work,
I think.
>
> SQLite is interesting because it's very widely used, runs on iPhone and
> Android, is a single file, performs well, and is lightweight.
>
> The hard part is getting the database to write into .ODB files instead of
> what it natively uses?
Indeed a topic I did not look into. Actually how is it done today? I do
not believe HSSQLDB has a capability for that, right?
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org
Re: [discussion] future embedded DB in AOO
Posted by Damjan Jovanovic <da...@apache.org>.
On Mon, Dec 7, 2020 at 8:38 PM Peter Kovacs <pe...@apache.org> wrote:
> Hi all,
>
> I would like to know how you guys feel we should move on with our base
> Database. Our current strategy is a small sized embedded DB using HSQLDB
> (BSD). There are 2 other options in the same weight class, that is H2
> (EPL1 or MPL2) and Apache Derby (AL2).
>
> Something like SQLLite (PublicDomain) could be technical interesting,
> but I think their licensing is not so appealing.
>
> We could also consider firebird (MPL) like LO, but I have the impression
> that this is not a lightweight DB, but featurerich. And I wonder if this
> is really something we should provide. I would rather develop a way to
> make it easy to switch the DB when a Project grows.
>
> What are your thoughts?
>
>
Whatever we do, we have to provide backward compatibility for HSQLDB.
I am more interested in something like Apache Calcite, which would let us
do SQL queries spanning multiple database backends, than in yet another
database backend.
SQLite is interesting because it's very widely used, runs on iPhone and
Android, is a single file, performs well, and is lightweight.
The hard part is getting the database to write into .ODB files instead of
what it natively uses?
Damjan
[discussion] future embedded DB in AOO
Posted by Mechtilde <oo...@mechtilde.de>.
Hello Peter,
Am 07.12.20 um 19:38 schrieb Peter Kovacs:
> Hi all,
>
> I would like to know how you guys feel we should move on with our base
> Database. Our current strategy is a small sized embedded DB using HSQLDB
> (BSD). There are 2 other options in the same weight class, that is H2
> (EPL1 or MPL2) and Apache Derby (AL2).
I prefer H2by several reasens:
1. H2 is the continued deelopment of HSQLDB
2. H2 is also used in other applications like Hibiscus (and JVerein ;-))
3. I know also some properity software which use it as an embedded DB
kind regards
Mechtilde
>
> Something like SQLLite (PublicDomain) could be technical interesting,
> but I think their licensing is not so appealing.
>
> We could also consider firebird (MPL) like LO, but I have the impression
> that this is not a lightweight DB, but featurerich. And I wonder if this
> is really something we should provide. I would rather develop a way to
> make it easy to switch the DB when a Project grows.
>
> What are your thoughts?
>
>
> All the Best
>
> Peter
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
> For additional commands, e-mail: dev-help@openoffice.apache.org
>
--
Mechtilde Stehmann
## Apache OpenOffice
## Freie Office Suite für Linux, MacOSX, Windows und OS/2
## Debian Developer
## PGP encryption welcome
## F0E3 7F3D C87A 4998 2899 39E7 F287 7BBA 141A AD7F