AOO Downloads, false-positive and Beyond

[Roberto Galoppini <rg...@geek.net>]

Hi all,

 as you might now SourceForge is serving the vast majority of AOO binaries
downloads, and we provide download stats by country, Operating System,
Browser, and traffic source. Some of you are familiar with our stats pages,
while others use our APIs.

We do have spam detection enabled to identify false-positive traffic like
bots, and I wish to share some insights of what happened recently on this
front.

We noticed that Russia was the highest download country, something that was
hard to explain. The popularity was limited to /localized/ru/3.4.0/
Apache_OpenOffice_incubating_3.4.0_Win_x86_install_ru.exe

We looked at the raw download logs for that file, and saw a lot of
downloads from user-agent "Download Master".  Apparently it is popular in
Russia, and apparently it starts hundreds of simultaneous downloads at
once.  Our download stats system does have some logic to prevent double
counting this type of traffic, but it didn't exclude all of the duplicate
downloads, so the result was still high.  We've updated our download stats
logic to correct this, and then reprocessed the raw logs from 2012-05-08 to
present, to update the stats.

Beyond bringing our ability to provide reliable stats, I wish to throw some
new ideas about how we can help Apache OpenOffice to grow:

 a. We could provide intelligence on which projects were downloaded with
Open Office within a week.
b. We could cross-merch Apache OpenOffice project with other projects
c. We have community management and Internet Marketing to support the
community

We've already run a 250k impressions campaign through our media channels,
and we plan to run more.
Our community growth hacker and Apache member Rich Bowen has covered Apache
OpenOffice both on feathercast and SourceForge blog, and also here we plan
to do more videos and interviews.

Roberto

-- 
====
This e- mail message is intended only for the named recipient(s) above. It 
may contain confidential and privileged information. If you are not the 
intended recipient you are hereby notified that any dissemination, 
distribution or copying of this e-mail and any attachment(s) is strictly 
prohibited. If you have received this e-mail in error, please immediately 
notify the sender by replying to this e-mail and delete the message and any 
attachment(s) from your system. Thank you.

Re: AOO Downloads, false-positive and Beyond

[Louis Suárez-Potts <lu...@gmail.com>]

On 2012-05-18, at 12:15 , Roberto Galoppini wrote:

>> 
>> 
>> As well, wouldn't it be great that over summer we do enough so that when
>> school starts again here in the Northern Hemisphere, students and faculty
>> can actually use something that is all about working together for a better
>> place?
>> 
> 
> Yea, any specific idea in mind?

Yes. Many. 

What I'd suggest is not to target a particular regional market but a logical one. In this case, let's look at, oh, say, the Portuguese deployment of the Magellan education netbooks. As you recall, earlier instances had Win/MS Office in one partition, Lin/OOo in the other. That implicitly and immediately deprecated free software, as there was little reason to use it.

As Paulo has told me, though, things have changed. But the basic elements are the same: gov't. promoted software (now more free) and devices that are inexpensive enough (esp. aftter support) so that they can be widely deployed. Portugal is small—about the size of a Beijing suburb, at least in terms of population (10M) and Brazil is big, nearly 20 times that size. There is also Angola and other former Lusophone polities around the world. 

My argument is to promote a solution that is, at first, software focused. 

* AOO in Portuguese (Br/Pt) with local, when possible, support. (If it is not there, we start it or use global options until it is—and it comes into being because the market is or ought to be obvious enough.)

* Devices: Intel sided publicly with LibreOffice but earlier had wanted to work with OOo. Frankly, I doubt Intel cares one way or another which flavour is the best but simply wants a suite that has humongous global usage and also crucial community and industrial momentum. 

We have both.

> So, we contact Intel. I know the people at the Tizen (formerly Meego) project, and I'm sure others here do, too. Tizen is essentially a mobilized Linux, I believe, and if Intel has worked to get LO on it—I don't know—then whatever it has done would probably work with AOO.

The point is to have a package: software on hardware. It need not be mobile per se. The education laptops, too, may not—I don't think they do—use Tizen. 

The most important point, as I see it, is to have a product that is ready for education—students and teachers—and only secondarily, the hardware. I put it this way, even though I obviously believe in the integrated ensemble, just because the hardware element is so susceptible to change. As well, hardware decisions are usually made more than a year in advance; I don't know when software decisions are made in general but I guess is that the cycle is a little shorter.

I think developing this general idea is needed, of course, and to focus on PT is not essential. However, my reason for doing so rests upon what had already been done by Portugal's prior gov't., by the PT users/OOo group, by the wide scale and exciting deployments in Brazil, by Angola's earlier interest in OOo and by Intel's live interest in producing a product that can have global popularity.

But all the players here could be replaced by others, of course. And the language could be, I don't know, Italian :-) or even the English they speak up here in Canada, eh?

Cheers,

Louis
> 
> Roberto
> 

Re: AOO Downloads, false-positive and Beyond

[Louis Suárez-Potts <lu...@gmail.com>]

On 2012-05-18, at 12:15 , Roberto Galoppini wrote:

>> 
>> 
>> As well, wouldn't it be great that over summer we do enough so that when
>> school starts again here in the Northern Hemisphere, students and faculty
>> can actually use something that is all about working together for a better
>> place?
>> 
> 
> Yea, any specific idea in mind?

Yes. Many. 

What I'd suggest is not to target a particular regional market but a logical one. In this case, let's look at, oh, say, the Portuguese deployment of the Magellan education netbooks. As you recall, earlier instances had Win/MS Office in one partition, Lin/OOo in the other. That implicitly and immediately deprecated free software, as there was little reason to use it.

As Paulo has told me, though, things have changed. But the basic elements are the same: gov't. promoted software (now more free) and devices that are inexpensive enough (esp. aftter support) so that they can be widely deployed. Portugal is small—about the size of a Beijing suburb, at least in terms of population (10M) and Brazil is big, nearly 20 times that size. There is also Angola and other former Lusophone polities around the world. 

My argument is to promote a solution that is, at first, software focused. 

* AOO in Portuguese (Br/Pt) with local, when possible, support. (If it is not there, we start it or use global options until it is—and it comes into being because the market is or ought to be obvious enough.)

* Devices: Intel sided publicly with LibreOffice but earlier had wanted to work with OOo. Frankly, I doubt Intel cares one way or another which flavour is the best but simply wants a suite that has humongous global usage and also crucial community and industrial momentum. 

We have both.

> So, we contact Intel. I know the people at the Tizen (formerly Meego) project, and I'm sure others here do, too. Tizen is essentially a mobilized Linux, I believe, and if Intel has worked to get LO on it—I don't know—then whatever it has done would probably work with AOO.

The point is to have a package: software on hardware. It need not be mobile per se. The education laptops, too, may not—I don't think they do—use Tizen. 

The most important point, as I see it, is to have a product that is ready for education—students and teachers—and only secondarily, the hardware. I put it this way, even though I obviously believe in the integrated ensemble, just because the hardware element is so susceptible to change. As well, hardware decisions are usually made more than a year in advance; I don't know when software decisions are made in general but I guess is that the cycle is a little shorter.

I think developing this general idea is needed, of course, and to focus on PT is not essential. However, my reason for doing so rests upon what had already been done by Portugal's prior gov't., by the PT users/OOo group, by the wide scale and exciting deployments in Brazil, by Angola's earlier interest in OOo and by Intel's live interest in producing a product that can have global popularity.

But all the players here could be replaced by others, of course. And the language could be, I don't know, Italian :-) or even the English they speak up here in Canada, eh?

Cheers,

Louis
> 
> Roberto
> 

Re: AOO Downloads, false-positive and Beyond

[Roberto Galoppini <rg...@geek.net>]

On Wed, May 16, 2012 at 4:25 AM, Louis Suárez-Potts <lu...@gmail.com>wrote:

>
> On 2012-05-15, at 14:49 , Roberto Galoppini wrote:
>
> > Hi all,
> >
> > as you might now SourceForge is serving the vast majority of AOO binaries
> > downloads, and we provide download stats by country, Operating System,
> > Browser, and traffic source. Some of you are familiar with our stats
> pages,
> > while others use our APIs.
> >
> > We do have spam detection enabled to identify false-positive traffic like
> > bots, and I wish to share some insights of what happened recently on this
> > front.
> >
> > We noticed that Russia was the highest download country, something that
> was
> > hard to explain. The popularity was limited to /localized/ru/3.4.0/
> > Apache_OpenOffice_incubating_3.4.0_Win_x86_install_ru.exe
> >
> > We looked at the raw download logs for that file, and saw a lot of
> > downloads from user-agent "Download Master".  Apparently it is popular in
> > Russia, and apparently it starts hundreds of simultaneous downloads at
> > once.  Our download stats system does have some logic to prevent double
> > counting this type of traffic, but it didn't exclude all of the duplicate
> > downloads, so the result was still high.  We've updated our download
> stats
> > logic to correct this, and then reprocessed the raw logs from 2012-05-08
> to
> > present, to update the stats.
> >
> > Beyond bringing our ability to provide reliable stats, I wish to throw
> some
> > new ideas about how we can help Apache OpenOffice to grow:
> >
> > a. We could provide intelligence on which projects were downloaded with
> > Open Office within a week.
> > b. We could cross-merch Apache OpenOffice project with other projects
> > c. We have community management and Internet Marketing to support the
> > community
>
> I'm in favour, as you probably can guess--I strongly promoted OOo both as
> a binary for users and as a source project for developers (considerable
> overlap)--but do have simple procedural questions, starting with "we"?
>
You mean, I'd hope, those who simply want to do it?



In this case 'we' was simply SourceForge.


As we encountered with the OOo Marketing Project, good ideas and intentions
> can quickly get lost in community cacophony: more noise than signal.
>
> What we discovered was that focusing on particular drives and engaging
> those who would be able to carry them out, long term and without undue
> stress to their regular lives (this is all volunteer), helped. What I
> further discovered and tried as much as possible to arrange was the support
> & coordination of small, medium and even large businesses and public sector
> entities. For instance, a company may have an extension that adds value to
> AOO and which, by its use, adds huge marketing value to their company and
> product. I received *a lot* of such requests from companies, and I would
> like to reacquaint myself with them and they with us, but it takes time.
>
> A preliminary list of organizations that were using OOo and probably are
> interested in AOO can be found at
> http://wiki.services.openoffice.org/wiki/Major_OpenOffice.org_Deployments
>
> The thing that I noted repeatedly was that many organizations, esp. public
> sector (and also not a small number of individuals coming from a Windows or
> Mac background) refused or were reluctant to download the product without
> professional support. The user forums worked great but in the case of
> public sectors and also companies, they wanted professional support, as
> they were used to getting (and paying for). This does not mean we must wait
> for the horse to be hitched to this wagon, not by any means. And I'm
> working on rekindling those who *were* providing that support. (Besides
> Sun/Oracle, there were actually quite a few. Some can still be found from
> http://support.openoffice.org/)
>
>
> >
> > We've already run a 250k impressions campaign through our media channels,
> > and we plan to run more.
> > Our community growth hacker and Apache member Rich Bowen has covered
> Apache
> > OpenOffice both on feathercast and SourceForge blog, and also here we
> plan
> > to do more videos and interviews.
> >
>
>
> > Roberto
> All in all, thanks, Roberto! I would suggest an IRC meeting with an agenda
> to start coordinating activities. I also see some implicit milestones.
> These include drawing attention to what is here, what is coming and how
> people can use it and contribute to it--without thinking about the Cloud,
> or cost. And if they must, that there are options there, too.
>

Apparently ooo-dev is still the best place to coordinate marketing-related
efforts.

I'm working on how we can pass the request of help to our developers'
audience, and possibly find some translators and/or devs.



>
> As well, wouldn't it be great that over summer we do enough so that when
> school starts again here in the Northern Hemisphere, students and faculty
> can actually use something that is all about working together for a better
> place?
>

Yea, any specific idea in mind?

Roberto


>
> Ciao
> Louis

-- 
====
This e- mail message is intended only for the named recipient(s) above. It 
may contain confidential and privileged information. If you are not the 
intended recipient you are hereby notified that any dissemination, 
distribution or copying of this e-mail and any attachment(s) is strictly 
prohibited. If you have received this e-mail in error, please immediately 
notify the sender by replying to this e-mail and delete the message and any 
attachment(s) from your system. Thank you.

Re: AOO Downloads, false-positive and Beyond

[Roberto Galoppini <rg...@geek.net>]

On Wed, May 16, 2012 at 4:25 AM, Louis Suárez-Potts <lu...@gmail.com>wrote:

>
> On 2012-05-15, at 14:49 , Roberto Galoppini wrote:
>
> > Hi all,
> >
> > as you might now SourceForge is serving the vast majority of AOO binaries
> > downloads, and we provide download stats by country, Operating System,
> > Browser, and traffic source. Some of you are familiar with our stats
> pages,
> > while others use our APIs.
> >
> > We do have spam detection enabled to identify false-positive traffic like
> > bots, and I wish to share some insights of what happened recently on this
> > front.
> >
> > We noticed that Russia was the highest download country, something that
> was
> > hard to explain. The popularity was limited to /localized/ru/3.4.0/
> > Apache_OpenOffice_incubating_3.4.0_Win_x86_install_ru.exe
> >
> > We looked at the raw download logs for that file, and saw a lot of
> > downloads from user-agent "Download Master".  Apparently it is popular in
> > Russia, and apparently it starts hundreds of simultaneous downloads at
> > once.  Our download stats system does have some logic to prevent double
> > counting this type of traffic, but it didn't exclude all of the duplicate
> > downloads, so the result was still high.  We've updated our download
> stats
> > logic to correct this, and then reprocessed the raw logs from 2012-05-08
> to
> > present, to update the stats.
> >
> > Beyond bringing our ability to provide reliable stats, I wish to throw
> some
> > new ideas about how we can help Apache OpenOffice to grow:
> >
> > a. We could provide intelligence on which projects were downloaded with
> > Open Office within a week.
> > b. We could cross-merch Apache OpenOffice project with other projects
> > c. We have community management and Internet Marketing to support the
> > community
>
> I'm in favour, as you probably can guess--I strongly promoted OOo both as
> a binary for users and as a source project for developers (considerable
> overlap)--but do have simple procedural questions, starting with "we"?
>
You mean, I'd hope, those who simply want to do it?



In this case 'we' was simply SourceForge.


As we encountered with the OOo Marketing Project, good ideas and intentions
> can quickly get lost in community cacophony: more noise than signal.
>
> What we discovered was that focusing on particular drives and engaging
> those who would be able to carry them out, long term and without undue
> stress to their regular lives (this is all volunteer), helped. What I
> further discovered and tried as much as possible to arrange was the support
> & coordination of small, medium and even large businesses and public sector
> entities. For instance, a company may have an extension that adds value to
> AOO and which, by its use, adds huge marketing value to their company and
> product. I received *a lot* of such requests from companies, and I would
> like to reacquaint myself with them and they with us, but it takes time.
>
> A preliminary list of organizations that were using OOo and probably are
> interested in AOO can be found at
> http://wiki.services.openoffice.org/wiki/Major_OpenOffice.org_Deployments
>
> The thing that I noted repeatedly was that many organizations, esp. public
> sector (and also not a small number of individuals coming from a Windows or
> Mac background) refused or were reluctant to download the product without
> professional support. The user forums worked great but in the case of
> public sectors and also companies, they wanted professional support, as
> they were used to getting (and paying for). This does not mean we must wait
> for the horse to be hitched to this wagon, not by any means. And I'm
> working on rekindling those who *were* providing that support. (Besides
> Sun/Oracle, there were actually quite a few. Some can still be found from
> http://support.openoffice.org/)
>
>
> >
> > We've already run a 250k impressions campaign through our media channels,
> > and we plan to run more.
> > Our community growth hacker and Apache member Rich Bowen has covered
> Apache
> > OpenOffice both on feathercast and SourceForge blog, and also here we
> plan
> > to do more videos and interviews.
> >
>
>
> > Roberto
> All in all, thanks, Roberto! I would suggest an IRC meeting with an agenda
> to start coordinating activities. I also see some implicit milestones.
> These include drawing attention to what is here, what is coming and how
> people can use it and contribute to it--without thinking about the Cloud,
> or cost. And if they must, that there are options there, too.
>

Apparently ooo-dev is still the best place to coordinate marketing-related
efforts.

I'm working on how we can pass the request of help to our developers'
audience, and possibly find some translators and/or devs.



>
> As well, wouldn't it be great that over summer we do enough so that when
> school starts again here in the Northern Hemisphere, students and faculty
> can actually use something that is all about working together for a better
> place?
>

Yea, any specific idea in mind?

Roberto


>
> Ciao
> Louis

-- 
====
This e- mail message is intended only for the named recipient(s) above. It 
may contain confidential and privileged information. If you are not the 
intended recipient you are hereby notified that any dissemination, 
distribution or copying of this e-mail and any attachment(s) is strictly 
prohibited. If you have received this e-mail in error, please immediately 
notify the sender by replying to this e-mail and delete the message and any 
attachment(s) from your system. Thank you.

Re: AOO Downloads, false-positive and Beyond

[Louis Suárez-Potts <lu...@gmail.com>]

On 2012-05-15, at 14:49 , Roberto Galoppini wrote:

> Hi all,
> 
> as you might now SourceForge is serving the vast majority of AOO binaries
> downloads, and we provide download stats by country, Operating System,
> Browser, and traffic source. Some of you are familiar with our stats pages,
> while others use our APIs.
> 
> We do have spam detection enabled to identify false-positive traffic like
> bots, and I wish to share some insights of what happened recently on this
> front.
> 
> We noticed that Russia was the highest download country, something that was
> hard to explain. The popularity was limited to /localized/ru/3.4.0/
> Apache_OpenOffice_incubating_3.4.0_Win_x86_install_ru.exe
> 
> We looked at the raw download logs for that file, and saw a lot of
> downloads from user-agent "Download Master".  Apparently it is popular in
> Russia, and apparently it starts hundreds of simultaneous downloads at
> once.  Our download stats system does have some logic to prevent double
> counting this type of traffic, but it didn't exclude all of the duplicate
> downloads, so the result was still high.  We've updated our download stats
> logic to correct this, and then reprocessed the raw logs from 2012-05-08 to
> present, to update the stats.
> 
> Beyond bringing our ability to provide reliable stats, I wish to throw some
> new ideas about how we can help Apache OpenOffice to grow:
> 
> a. We could provide intelligence on which projects were downloaded with
> Open Office within a week.
> b. We could cross-merch Apache OpenOffice project with other projects
> c. We have community management and Internet Marketing to support the
> community

I'm in favour, as you probably can guess--I strongly promoted OOo both as a binary for users and as a source project for developers (considerable overlap)--but do have simple procedural questions, starting with "we"? You mean, I'd hope, those who simply want to do it? As we encountered with the OOo Marketing Project, good ideas and intentions can quickly get lost in community cacophony: more noise than signal. 

What we discovered was that focusing on particular drives and engaging those who would be able to carry them out, long term and without undue stress to their regular lives (this is all volunteer), helped. What I further discovered and tried as much as possible to arrange was the support & coordination of small, medium and even large businesses and public sector entities. For instance, a company may have an extension that adds value to AOO and which, by its use, adds huge marketing value to their company and product. I received *a lot* of such requests from companies, and I would like to reacquaint myself with them and they with us, but it takes time.

A preliminary list of organizations that were using OOo and probably are interested in AOO can be found at http://wiki.services.openoffice.org/wiki/Major_OpenOffice.org_Deployments

The thing that I noted repeatedly was that many organizations, esp. public sector (and also not a small number of individuals coming from a Windows or Mac background) refused or were reluctant to download the product without professional support. The user forums worked great but in the case of public sectors and also companies, they wanted professional support, as they were used to getting (and paying for). This does not mean we must wait for the horse to be hitched to this wagon, not by any means. And I'm working on rekindling those who *were* providing that support. (Besides Sun/Oracle, there were actually quite a few. Some can still be found from http://support.openoffice.org/)


> 
> We've already run a 250k impressions campaign through our media channels,
> and we plan to run more.
> Our community growth hacker and Apache member Rich Bowen has covered Apache
> OpenOffice both on feathercast and SourceForge blog, and also here we plan
> to do more videos and interviews.
> 


> Roberto
All in all, thanks, Roberto! I would suggest an IRC meeting with an agenda to start coordinating activities. I also see some implicit milestones. These include drawing attention to what is here, what is coming and how people can use it and contribute to it--without thinking about the Cloud, or cost. And if they must, that there are options there, too.

As well, wouldn't it be great that over summer we do enough so that when school starts again here in the Northern Hemisphere, students and faculty can actually use something that is all about working together for a better place?

Ciao
Louis

Re: AOO Downloads, false-positive and Beyond

[Louis Suárez-Potts <lu...@gmail.com>]

On 2012-05-15, at 14:49 , Roberto Galoppini wrote:

> Hi all,
> 
> as you might now SourceForge is serving the vast majority of AOO binaries
> downloads, and we provide download stats by country, Operating System,
> Browser, and traffic source. Some of you are familiar with our stats pages,
> while others use our APIs.
> 
> We do have spam detection enabled to identify false-positive traffic like
> bots, and I wish to share some insights of what happened recently on this
> front.
> 
> We noticed that Russia was the highest download country, something that was
> hard to explain. The popularity was limited to /localized/ru/3.4.0/
> Apache_OpenOffice_incubating_3.4.0_Win_x86_install_ru.exe
> 
> We looked at the raw download logs for that file, and saw a lot of
> downloads from user-agent "Download Master".  Apparently it is popular in
> Russia, and apparently it starts hundreds of simultaneous downloads at
> once.  Our download stats system does have some logic to prevent double
> counting this type of traffic, but it didn't exclude all of the duplicate
> downloads, so the result was still high.  We've updated our download stats
> logic to correct this, and then reprocessed the raw logs from 2012-05-08 to
> present, to update the stats.
> 
> Beyond bringing our ability to provide reliable stats, I wish to throw some
> new ideas about how we can help Apache OpenOffice to grow:
> 
> a. We could provide intelligence on which projects were downloaded with
> Open Office within a week.
> b. We could cross-merch Apache OpenOffice project with other projects
> c. We have community management and Internet Marketing to support the
> community

I'm in favour, as you probably can guess--I strongly promoted OOo both as a binary for users and as a source project for developers (considerable overlap)--but do have simple procedural questions, starting with "we"? You mean, I'd hope, those who simply want to do it? As we encountered with the OOo Marketing Project, good ideas and intentions can quickly get lost in community cacophony: more noise than signal. 

What we discovered was that focusing on particular drives and engaging those who would be able to carry them out, long term and without undue stress to their regular lives (this is all volunteer), helped. What I further discovered and tried as much as possible to arrange was the support & coordination of small, medium and even large businesses and public sector entities. For instance, a company may have an extension that adds value to AOO and which, by its use, adds huge marketing value to their company and product. I received *a lot* of such requests from companies, and I would like to reacquaint myself with them and they with us, but it takes time.

A preliminary list of organizations that were using OOo and probably are interested in AOO can be found at http://wiki.services.openoffice.org/wiki/Major_OpenOffice.org_Deployments

The thing that I noted repeatedly was that many organizations, esp. public sector (and also not a small number of individuals coming from a Windows or Mac background) refused or were reluctant to download the product without professional support. The user forums worked great but in the case of public sectors and also companies, they wanted professional support, as they were used to getting (and paying for). This does not mean we must wait for the horse to be hitched to this wagon, not by any means. And I'm working on rekindling those who *were* providing that support. (Besides Sun/Oracle, there were actually quite a few. Some can still be found from http://support.openoffice.org/)


> 
> We've already run a 250k impressions campaign through our media channels,
> and we plan to run more.
> Our community growth hacker and Apache member Rich Bowen has covered Apache
> OpenOffice both on feathercast and SourceForge blog, and also here we plan
> to do more videos and interviews.
> 


> Roberto
All in all, thanks, Roberto! I would suggest an IRC meeting with an agenda to start coordinating activities. I also see some implicit milestones. These include drawing attention to what is here, what is coming and how people can use it and contribute to it--without thinking about the Cloud, or cost. And if they must, that there are options there, too.

As well, wouldn't it be great that over summer we do enough so that when school starts again here in the Northern Hemisphere, students and faculty can actually use something that is all about working together for a better place?

Ciao
Louis