You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@poi.apache.org by Ryan Ackley <sa...@cfl.rr.com> on 2003/05/19 02:51:22 UTC
HDF
I wanted to ask for some feedback from POI committers...
Lately, there has been a lot of interest from people who want to contribute
to HDF. People have been submitting patches to code that is in scratchpad.
In some ways I think that the code I originally contributed has slowed down
the project. Some potential contributors may have taken one look at the
source code and been like "HOLY $&!@%$@ CRAP! thats ugly". I say that
because anytime someone wants to contribute, I tell them "study the spec and
look at the source code". I never hear from them again :-)
Bob Otterberg emailed me a High-Level design for an object model. I have
come up with a low-level design for reading AND writing the main data
structures. These are attached as a .zargo file. Use Poseidon UML
(http://www.gentleware.com) (free version) to open it. It is ArgoUML w/o all
the bugs. I think it will be easier to get other people involved in
developing HDF if we get rid of the legacy baggage (my old code). Also, the
quality will be better because we have designed it from the beginning to
serve the needs of the community as opposed to the needs of Ryan. Of course,
my old code will be a valuable reference for this new development effort
Two things:
1) Bob and I want your feedback on the design. If you don't understand
something, please ask.
2) How do I get people involved and keep them interested? Its a little
easier managing people when they aren't working on a volunteer basis. When
John Smith from internet land posts a message to the list asking how they
can contribute to HDF. Do I assign him a task that may never be completed?
Or do I tell them to figure out what needs to be done?
Thanks,
Ryan
Re: HDF
Posted by "Andrew C. Oliver" <ac...@apache.org>.
I offer this: http://fishbowl.pastiche.org/archives/001331.html as
supporting information ;-)
-Andy
On 5/18/03 6:51 PM, "Ryan Ackley" <sa...@cfl.rr.com> wrote:
> I wanted to ask for some feedback from POI committers...
>
> Lately, there has been a lot of interest from people who want to contribute
> to HDF. People have been submitting patches to code that is in scratchpad.
> In some ways I think that the code I originally contributed has slowed down
> the project. Some potential contributors may have taken one look at the
> source code and been like "HOLY $&!@%$@ CRAP! thats ugly". I say that
> because anytime someone wants to contribute, I tell them "study the spec and
> look at the source code". I never hear from them again :-)
>
> Bob Otterberg emailed me a High-Level design for an object model. I have
> come up with a low-level design for reading AND writing the main data
> structures. These are attached as a .zargo file. Use Poseidon UML
> (http://www.gentleware.com) (free version) to open it. It is ArgoUML w/o all
> the bugs. I think it will be easier to get other people involved in
> developing HDF if we get rid of the legacy baggage (my old code). Also, the
> quality will be better because we have designed it from the beginning to
> serve the needs of the community as opposed to the needs of Ryan. Of course,
> my old code will be a valuable reference for this new development effort
>
> Two things:
>
> 1) Bob and I want your feedback on the design. If you don't understand
> something, please ask.
>
> 2) How do I get people involved and keep them interested? Its a little
> easier managing people when they aren't working on a volunteer basis. When
> John Smith from internet land posts a message to the list asking how they
> can contribute to HDF. Do I assign him a task that may never be completed?
> Or do I tell them to figure out what needs to be done?
>
> Thanks,
>
> Ryan
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: poi-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: poi-dev-help@jakarta.apache.org
--
Andrew C. Oliver
http://www.superlinksoftware.com/poi.jsp
Custom enhancements and Commercial Implementation for Jakarta POI
http://jakarta.apache.org/poi
For Java and Excel, Got POI?
Re: HDF
Posted by "Andrew C. Oliver" <ac...@apache.org>.
Spec schmeck. These are Microsoft file formats. Even a clear spec would be
convoluted. The file format is lame and convoluted. That¹s the fun. We
write tools, we write unit tests, and that¹s all the spec you need. But
what do I know ;-)
-Andy
On 5/24/03 7:29 PM, "Serge Huber2" <sh...@jahia.com> wrote:
>
> If I may comment on this discussion as somebody who tried to work with the
> existing code base...
>
> My problem wasn't as much the code as the actual spec. I've never seen
> something so convoluted or overly complicated in quite a while. I think
> that it would be a big help to structure the existing Word document spec in
> a more human-understandable way. When I was having trouble with something
> that wasn't working, I was looking at the spec, reading it, re-reading it,
> re-reading it again and again, and realized that I had to jump in a lot of
> places to figure out just one data structure (for example the piece tables
> is in my opinion very badly explained).
>
> Also, I suspect that the existing spec is not complete, and that there are
> things that have changed slightly (I noticed weird things going on with
> piece tables in Word XP documents or fast-saved documents).
>
> Anyway that's my two cents. I think the work that has already been done is
> nothing short of amazing, implementing a spec that complicated is really an
> achievement. If we can make it into something that more people understand
> well so much the better.
>
> Regards,
> Serge Huber.
>
> At 20:49 18.05.2003 -0600, you wrote:
>> Okay, I think getting RID of the "legacy code" is a bit drastic. It makes
>> far more sense to me to just start anew with a new abbreviation (we may have
>> inadvertently infringed on someone's trademark anyhow, though theoretically
>> they may have ceded it by not complaining up to now). Leave HDF out as an
>> old source base.
>>
>> You also are making IMHO a misinterpretation of opensource. The interest is
>> because there is code there and they try and use it and patch it as it
>> fails. What happens when you have 1-2 classes there and a big dream?
>> Nothing. Not one person will contribute (outside you two). Do you know how
>> many contributors Marc and I had before POI worked to any degree? 0. Nynt,
>> Kein, Cero.
>>
>> So my vote is leave it there and reap the benefit of the work that¹s going
>> on. Quite the opposite from your belief I think it will HELP you. Some
>> folks will come to the same dreadful conclusion and be interested in helping
>> with the new effort. You'll get patches to the old code that reveal
>> something needed in the new.
>>
>> Also the next iteration of our O'Reilly series is undergoing development
>> (we've had 2 onjava.com articles so far). I'd like to cover the Word
>> effort. Would one of you be interested in helping/doing-all-of writing that
>> part? No money, just exposure for POI (and yourself[-ves]). Here's hoping
>> I find a way to send this before next week...
>>
>> -Andy
>>
>> On 5/18/03 7:20 PM, "Shawn Laubach" <sh...@cox.net> wrote:
>>
>>> The attachment didn't make it (at least to me). I would be very interested
>>> in a copy. I'm trying to get back into POI development.
>>>
>>> Shawn
>>>
>>> -----Original Message-----
>>> From: Ryan Ackley [mailto:sackley@cfl.rr.com]
>>> Sent: Sunday, May 18, 2003 7:51 PM
>>> To: POI Developers List
>>> Cc: Bob Otterberg
>>> Subject: HDF
>>>
>>>
>>> I wanted to ask for some feedback from POI committers...
>>>
>>> Lately, there has been a lot of interest from people who want to contribute
>>> to HDF. People have been submitting patches to code that is in scratchpad.
>>> In some ways I think that the code I originally contributed has slowed down
>>> the project. Some potential contributors may have taken one look at the
>>> source code and been like "HOLY $&!@%$@ CRAP! thats ugly". I say that
>>> because anytime someone wants to contribute, I tell them "study the
>> spec and
>>> look at the source code". I never hear from them again :-)
>>>
>>> Bob Otterberg emailed me a High-Level design for an object model. I have
>>> come up with a low-level design for reading AND writing the main data
>>> structures. These are attached as a .zargo file. Use Poseidon UML
>>> (http://www.gentleware.com) (free version) to open it. It is ArgoUML
>> w/o all
>>> the bugs. I think it will be easier to get other people involved in
>>> developing HDF if we get rid of the legacy baggage (my old code). Also, the
>>> quality will be better because we have designed it from the beginning to
>>> serve the needs of the community as opposed to the needs of Ryan. Of
>> course,
>>> my old code will be a valuable reference for this new development effort
>>>
>>> Two things:
>>>
>>> 1) Bob and I want your feedback on the design. If you don't understand
>>> something, please ask.
>>>
>>> 2) How do I get people involved and keep them interested? Its a little
>>> easier managing people when they aren't working on a volunteer basis. When
>>> John Smith from internet land posts a message to the list asking how they
>>> can contribute to HDF. Do I assign him a task that may never be completed?
>>> Or do I tell them to figure out what needs to be done?
>>>
>>> Thanks,
>>>
>>> Ryan
>>>
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: poi-dev-unsubscribe@jakarta.apache.org
>>> For additional commands, e-mail: poi-dev-help@jakarta.apache.org
>>>
>>
>> --
>> Andrew C. Oliver
>> http://www.superlinksoftware.com/poi.jsp
>> Custom enhancements and Commercial Implementation for Jakarta POI
>>
>> http://jakarta.apache.org/poi
>> For Java and Excel, Got POI?
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: poi-dev-unsubscribe@jakarta.apache.org
>> For additional commands, e-mail: poi-dev-help@jakarta.apache.org
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: poi-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: poi-dev-help@jakarta.apache.org
>
--
Andrew C. Oliver
http://www.superlinksoftware.com/poi.jsp
Custom enhancements and Commercial Implementation for Jakarta POI
http://jakarta.apache.org/poi
For Java and Excel, Got POI?
Re: HDF
Posted by Serge Huber2 <sh...@jahia.com>.
If I may comment on this discussion as somebody who tried to work with the
existing code base...
My problem wasn't as much the code as the actual spec. I've never seen
something so convoluted or overly complicated in quite a while. I think
that it would be a big help to structure the existing Word document spec in
a more human-understandable way. When I was having trouble with something
that wasn't working, I was looking at the spec, reading it, re-reading it,
re-reading it again and again, and realized that I had to jump in a lot of
places to figure out just one data structure (for example the piece tables
is in my opinion very badly explained).
Also, I suspect that the existing spec is not complete, and that there are
things that have changed slightly (I noticed weird things going on with
piece tables in Word XP documents or fast-saved documents).
Anyway that's my two cents. I think the work that has already been done is
nothing short of amazing, implementing a spec that complicated is really an
achievement. If we can make it into something that more people understand
well so much the better.
Regards,
Serge Huber.
At 20:49 18.05.2003 -0600, you wrote:
>Okay, I think getting RID of the "legacy code" is a bit drastic. It makes
>far more sense to me to just start anew with a new abbreviation (we may have
>inadvertently infringed on someone's trademark anyhow, though theoretically
>they may have ceded it by not complaining up to now). Leave HDF out as an
>old source base.
>
>You also are making IMHO a misinterpretation of opensource. The interest is
>because there is code there and they try and use it and patch it as it
>fails. What happens when you have 1-2 classes there and a big dream?
>Nothing. Not one person will contribute (outside you two). Do you know how
>many contributors Marc and I had before POI worked to any degree? 0. Nynt,
>Kein, Cero.
>
>So my vote is leave it there and reap the benefit of the work that¹s going
>on. Quite the opposite from your belief I think it will HELP you. Some
>folks will come to the same dreadful conclusion and be interested in helping
>with the new effort. You'll get patches to the old code that reveal
>something needed in the new.
>
>Also the next iteration of our O'Reilly series is undergoing development
>(we've had 2 onjava.com articles so far). I'd like to cover the Word
>effort. Would one of you be interested in helping/doing-all-of writing that
>part? No money, just exposure for POI (and yourself[-ves]). Here's hoping
>I find a way to send this before next week...
>
>-Andy
>
>On 5/18/03 7:20 PM, "Shawn Laubach" <sh...@cox.net> wrote:
>
> > The attachment didn't make it (at least to me). I would be very interested
> > in a copy. I'm trying to get back into POI development.
> >
> > Shawn
> >
> > -----Original Message-----
> > From: Ryan Ackley [mailto:sackley@cfl.rr.com]
> > Sent: Sunday, May 18, 2003 7:51 PM
> > To: POI Developers List
> > Cc: Bob Otterberg
> > Subject: HDF
> >
> >
> > I wanted to ask for some feedback from POI committers...
> >
> > Lately, there has been a lot of interest from people who want to contribute
> > to HDF. People have been submitting patches to code that is in scratchpad.
> > In some ways I think that the code I originally contributed has slowed down
> > the project. Some potential contributors may have taken one look at the
> > source code and been like "HOLY $&!@%$@ CRAP! thats ugly". I say that
> > because anytime someone wants to contribute, I tell them "study the
> spec and
> > look at the source code". I never hear from them again :-)
> >
> > Bob Otterberg emailed me a High-Level design for an object model. I have
> > come up with a low-level design for reading AND writing the main data
> > structures. These are attached as a .zargo file. Use Poseidon UML
> > (http://www.gentleware.com) (free version) to open it. It is ArgoUML
> w/o all
> > the bugs. I think it will be easier to get other people involved in
> > developing HDF if we get rid of the legacy baggage (my old code). Also, the
> > quality will be better because we have designed it from the beginning to
> > serve the needs of the community as opposed to the needs of Ryan. Of
> course,
> > my old code will be a valuable reference for this new development effort
> >
> > Two things:
> >
> > 1) Bob and I want your feedback on the design. If you don't understand
> > something, please ask.
> >
> > 2) How do I get people involved and keep them interested? Its a little
> > easier managing people when they aren't working on a volunteer basis. When
> > John Smith from internet land posts a message to the list asking how they
> > can contribute to HDF. Do I assign him a task that may never be completed?
> > Or do I tell them to figure out what needs to be done?
> >
> > Thanks,
> >
> > Ryan
> >
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: poi-dev-unsubscribe@jakarta.apache.org
> > For additional commands, e-mail: poi-dev-help@jakarta.apache.org
> >
>
>--
>Andrew C. Oliver
>http://www.superlinksoftware.com/poi.jsp
>Custom enhancements and Commercial Implementation for Jakarta POI
>
>http://jakarta.apache.org/poi
>For Java and Excel, Got POI?
>
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: poi-dev-unsubscribe@jakarta.apache.org
>For additional commands, e-mail: poi-dev-help@jakarta.apache.org
RE: HDF
Posted by Michael Zalewski <za...@optonline.net>.
> We thought that it would be better on memory if we kept the structures
> compressed until needed. So a Section object would not be uncompressed
> until it was needed by the programmer. So the only uncompressed objects
are
> the ones the programmer has made direct calls on. Then, when a setter was
> called on a Section object it would create a compressed sprm (if needed)
in
> the SEPX. There is a memory-cpu trade off. Who knows, we may change our
> minds if it turns out to be real slow. We are trying to avoid an object
> model that is a total memory pig.
I have been working for some time now on a library to access Microsoft
Project
files. I use the same technique of not unpacking the structure until it is
accessed by the getter method. From my limited experience, any performance
price you pay is more than overwhelmed by the file I/O and by POIFS (the
Ole Structured File System).
And I think for large documents on small machines (small in POI terms means
less than 1 GB of memory it seems), I think the technique is actually faster
unless you can back up all the memory used by POI with real memory.
Re: HDF
Posted by Ryan Ackley <sa...@cfl.rr.com>.
> BTW, I was curious why you picked the SEPX to represent the section
> properties in the object model. I was thinking that the SEP structure
would
> be better with the read and write doing the SEPX structures?
Bob Otterberg and I came up with the object model design. One thing I didn't
put in the object model are the xml generated types. These are
CHPAbstractType, DOPAbstractType, FIBAbstractType, PAPAbstractType,
SEPAbstractType, TAPAbstractType, TCAbstractType. All of these line up with
the spec on variable names. For example, the Section class will extend
SEPAbstractType. In the Section class will be human understandable getters
and setters, not the funky names from the spec. This way most people won't
have to go to the spec to use HDF.
We thought that it would be better on memory if we kept the structures
compressed until needed. So a Section object would not be uncompressed
until it was needed by the programmer. So the only uncompressed objects are
the ones the programmer has made direct calls on. Then, when a setter was
called on a Section object it would create a compressed sprm (if needed) in
the SEPX. There is a memory-cpu trade off. Who knows, we may change our
minds if it turns out to be real slow. We are trying to avoid an object
model that is a total memory pig.
> I've started tinkering around with some code and getting where I
understand
> some of the format. If you've started on the code, I don't want to
> duplicate what you're doing. Would there be a way that could offload a
> thing or two easily that I can work on?
Since you asked, here is a list of things on my todo list in descending
order of priority:
Work on the above menitoned problem. Create classes that extend the xml
generated abstract types. Look at the spec and come up with human
understandable names for getters and setters. For example, a getter for the
right margin indent of a paragraph:
public int getRightMargin()
{
return super.getDxaRight();
}
Work on the setters so they will create a compressed sprm. I am working on
the compressed property classes such as SEPX, CHPX, PAPX.
The StyleSheet class and StyleDescription class need work. One problem I
have with the StyleSheet class is the way it uncompresses sprms. It is real
ugly. It consists of several giant switch statements. This is not a priority
because the way I do it now works fine.
I am currently working on low level reading and writing of the text,
section, character runs, and paragraphs. I will try to get something checked
in this weekend maybe.
Ryan Ackley
RE: HDF
Posted by Shawn Laubach <sh...@cox.net>.
I've started tinkering around with some code and getting where I understand
some of the format. If you've started on the code, I don't want to
duplicate what you're doing. Would there be a way that could offload a
thing or two easily that I can work on?
Shawn
BTW, I was curious why you picked the SEPX to represent the section
properties in the object model. I was thinking that the SEP structure would
be better with the read and write doing the SEPX structures?
-----Original Message-----
From: Ryan Ackley [mailto:sackley@cfl.rr.com]
Sent: Monday, May 26, 2003 7:51 PM
To: POI Developers List
Subject: Re: HDF
> So my vote is leave it there and reap the benefit of the work that¹s going
> on. Quite the opposite from your belief I think it will HELP you. Some
> folks will come to the same dreadful conclusion and be interested in
helping
> with the new effort. You'll get patches to the old code that reveal
> something needed in the new.
That makes sense.
> Also the next iteration of our O'Reilly series is undergoing development
> (we've had 2 onjava.com articles so far). I'd like to cover the Word
> effort. Would one of you be interested in helping/doing-all-of writing
that
> part? No money, just exposure for POI (and yourself[-ves]). Here's
hoping
> I find a way to send this before next week...
Development is coming along nicely. Surprisingly, I am able to re-use a lot
of the old stuff. I don't have it to a point that I can actually do
anything, so I am holding off checking anything in. Ideally, if you could
put them off for a couple of weeks or even a month, I think I could have a
version that can do read and write...
Ryan
---------------------------------------------------------------------
To unsubscribe, e-mail: poi-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: poi-dev-help@jakarta.apache.org
Re: HDF
Posted by Ryan Ackley <sa...@cfl.rr.com>.
> So my vote is leave it there and reap the benefit of the work that¹s going
> on. Quite the opposite from your belief I think it will HELP you. Some
> folks will come to the same dreadful conclusion and be interested in
helping
> with the new effort. You'll get patches to the old code that reveal
> something needed in the new.
That makes sense.
> Also the next iteration of our O'Reilly series is undergoing development
> (we've had 2 onjava.com articles so far). I'd like to cover the Word
> effort. Would one of you be interested in helping/doing-all-of writing
that
> part? No money, just exposure for POI (and yourself[-ves]). Here's
hoping
> I find a way to send this before next week...
Development is coming along nicely. Surprisingly, I am able to re-use a lot
of the old stuff. I don't have it to a point that I can actually do
anything, so I am holding off checking anything in. Ideally, if you could
put them off for a couple of weeks or even a month, I think I could have a
version that can do read and write...
Ryan
Re: HDF
Posted by "Andrew C. Oliver" <ac...@apache.org>.
Okay, I think getting RID of the "legacy code" is a bit drastic. It makes
far more sense to me to just start anew with a new abbreviation (we may have
inadvertently infringed on someone's trademark anyhow, though theoretically
they may have ceded it by not complaining up to now). Leave HDF out as an
old source base.
You also are making IMHO a misinterpretation of opensource. The interest is
because there is code there and they try and use it and patch it as it
fails. What happens when you have 1-2 classes there and a big dream?
Nothing. Not one person will contribute (outside you two). Do you know how
many contributors Marc and I had before POI worked to any degree? 0. Nynt,
Kein, Cero.
So my vote is leave it there and reap the benefit of the work that¹s going
on. Quite the opposite from your belief I think it will HELP you. Some
folks will come to the same dreadful conclusion and be interested in helping
with the new effort. You'll get patches to the old code that reveal
something needed in the new.
Also the next iteration of our O'Reilly series is undergoing development
(we've had 2 onjava.com articles so far). I'd like to cover the Word
effort. Would one of you be interested in helping/doing-all-of writing that
part? No money, just exposure for POI (and yourself[-ves]). Here's hoping
I find a way to send this before next week...
-Andy
On 5/18/03 7:20 PM, "Shawn Laubach" <sh...@cox.net> wrote:
> The attachment didn't make it (at least to me). I would be very interested
> in a copy. I'm trying to get back into POI development.
>
> Shawn
>
> -----Original Message-----
> From: Ryan Ackley [mailto:sackley@cfl.rr.com]
> Sent: Sunday, May 18, 2003 7:51 PM
> To: POI Developers List
> Cc: Bob Otterberg
> Subject: HDF
>
>
> I wanted to ask for some feedback from POI committers...
>
> Lately, there has been a lot of interest from people who want to contribute
> to HDF. People have been submitting patches to code that is in scratchpad.
> In some ways I think that the code I originally contributed has slowed down
> the project. Some potential contributors may have taken one look at the
> source code and been like "HOLY $&!@%$@ CRAP! thats ugly". I say that
> because anytime someone wants to contribute, I tell them "study the spec and
> look at the source code". I never hear from them again :-)
>
> Bob Otterberg emailed me a High-Level design for an object model. I have
> come up with a low-level design for reading AND writing the main data
> structures. These are attached as a .zargo file. Use Poseidon UML
> (http://www.gentleware.com) (free version) to open it. It is ArgoUML w/o all
> the bugs. I think it will be easier to get other people involved in
> developing HDF if we get rid of the legacy baggage (my old code). Also, the
> quality will be better because we have designed it from the beginning to
> serve the needs of the community as opposed to the needs of Ryan. Of course,
> my old code will be a valuable reference for this new development effort
>
> Two things:
>
> 1) Bob and I want your feedback on the design. If you don't understand
> something, please ask.
>
> 2) How do I get people involved and keep them interested? Its a little
> easier managing people when they aren't working on a volunteer basis. When
> John Smith from internet land posts a message to the list asking how they
> can contribute to HDF. Do I assign him a task that may never be completed?
> Or do I tell them to figure out what needs to be done?
>
> Thanks,
>
> Ryan
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: poi-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: poi-dev-help@jakarta.apache.org
>
--
Andrew C. Oliver
http://www.superlinksoftware.com/poi.jsp
Custom enhancements and Commercial Implementation for Jakarta POI
http://jakarta.apache.org/poi
For Java and Excel, Got POI?
RE: HDF
Posted by Shawn Laubach <sh...@cox.net>.
The attachment didn't make it (at least to me). I would be very interested
in a copy. I'm trying to get back into POI development.
Shawn
-----Original Message-----
From: Ryan Ackley [mailto:sackley@cfl.rr.com]
Sent: Sunday, May 18, 2003 7:51 PM
To: POI Developers List
Cc: Bob Otterberg
Subject: HDF
I wanted to ask for some feedback from POI committers...
Lately, there has been a lot of interest from people who want to contribute
to HDF. People have been submitting patches to code that is in scratchpad.
In some ways I think that the code I originally contributed has slowed down
the project. Some potential contributors may have taken one look at the
source code and been like "HOLY $&!@%$@ CRAP! thats ugly". I say that
because anytime someone wants to contribute, I tell them "study the spec and
look at the source code". I never hear from them again :-)
Bob Otterberg emailed me a High-Level design for an object model. I have
come up with a low-level design for reading AND writing the main data
structures. These are attached as a .zargo file. Use Poseidon UML
(http://www.gentleware.com) (free version) to open it. It is ArgoUML w/o all
the bugs. I think it will be easier to get other people involved in
developing HDF if we get rid of the legacy baggage (my old code). Also, the
quality will be better because we have designed it from the beginning to
serve the needs of the community as opposed to the needs of Ryan. Of course,
my old code will be a valuable reference for this new development effort
Two things:
1) Bob and I want your feedback on the design. If you don't understand
something, please ask.
2) How do I get people involved and keep them interested? Its a little
easier managing people when they aren't working on a volunteer basis. When
John Smith from internet land posts a message to the list asking how they
can contribute to HDF. Do I assign him a task that may never be completed?
Or do I tell them to figure out what needs to be done?
Thanks,
Ryan
Re: HDF
Posted by Rainer Klute <ra...@gmx.de>.
> At 10:51 AM 19/05/2003, Glen wrote:
> I wouldn't like to see HDF go away as we do have at least a few real
> people
> using it. If you really feel that the codebase is beyond repair perhaps a
> separate package can be created (while leaving the existing codebase in
> place). If the new codebase advances beyond the capabilities of the old
> one then that would be the time to consider permanently removing it. What
> do you think?
I like Glen's approach. I already downloaded Poseidon and will try to have a
look at the design later.
Best regards
Rainer Klute
--
RAINER KLUTE IT-CONSULTING GMBH
Dipl.-Inform.
Rainer Klute klute@rainer-klute.de
Körner Grund 24 Telefon: +49 172 2324824
D-44143 Dortmund Telefax: +49 231 5349423
+++ GMX - Mail, Messaging & more http://www.gmx.net +++
Bitte lächeln! Fotogalerie online mit GMX ohne eigene Homepage!
Re: HDF
Posted by Glen Stampoultzis <gs...@iinet.net.au>.
At 10:51 AM 19/05/2003, you wrote:
>I wanted to ask for some feedback from POI committers...
>
>Lately, there has been a lot of interest from people who want to contribute
>to HDF. People have been submitting patches to code that is in scratchpad.
>In some ways I think that the code I originally contributed has slowed down
>the project. Some potential contributors may have taken one look at the
>source code and been like "HOLY $&!@%$@ CRAP! thats ugly". I say that
>because anytime someone wants to contribute, I tell them "study the spec and
>look at the source code". I never hear from them again :-)
>
>Bob Otterberg emailed me a High-Level design for an object model. I have
>come up with a low-level design for reading AND writing the main data
>structures. These are attached as a .zargo file. Use Poseidon UML
>(http://www.gentleware.com) (free version) to open it. It is ArgoUML w/o all
>the bugs. I think it will be easier to get other people involved in
>developing HDF if we get rid of the legacy baggage (my old code). Also, the
>quality will be better because we have designed it from the beginning to
>serve the needs of the community as opposed to the needs of Ryan. Of course,
>my old code will be a valuable reference for this new development effort
>
>Two things:
>
>1) Bob and I want your feedback on the design. If you don't understand
>something, please ask.
>
>2) How do I get people involved and keep them interested? Its a little
>easier managing people when they aren't working on a volunteer basis. When
>John Smith from internet land posts a message to the list asking how they
>can contribute to HDF. Do I assign him a task that may never be completed?
>Or do I tell them to figure out what needs to be done?
I think it's easier to get people interested in an existing codebase than
to try and create a new codebase from scratch.
I wouldn't like to see HDF go away as we do have at least a few real people
using it. If you really feel that the codebase is beyond repair perhaps a
separate package can be created (while leaving the existing codebase in
place). If the new codebase advances beyond the capabilities of the old
one then that would be the time to consider permanently removing it. What
do you think?
Regards,
Glen