You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Ben Collins-Sussman <su...@red-bean.com> on 2009/04/16 14:44:02 UTC

interesting statistics

googlecode gets a huge amount of svn traffic.  I was looking at log
statistics, specifically at aggregated user-agent: strings.  Here's
the relative popularity of different svn clients:


cmdline  65%
SvnKit   15%
Tortoise  8%
SharpSvn  1%
other    11%  (too many useragent micro-variants to count)

cmdline breakdown:

1.6.x   5%
1.5.x   50%
1.4.x   30%
older   15%

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748379

Re: interesting statistics

Posted by Mark Phippard <ma...@gmail.com>.
On Thu, Apr 16, 2009 at 11:14 AM, Greg Stein <gs...@gmail.com> wrote:
> Wouldn't tools embedding our libraries appear as the cmdline?

Yes.  Although since 1.5 we expose an ability to customize the agent
string (which TortoiseSVN now uses).

-- 
Thanks

Mark Phippard
http://markphip.blogspot.com/

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748789

Re: interesting statistics

Posted by Greg Stein <gs...@gmail.com>.
Wouldn't tools embedding our libraries appear as the cmdline?

On Thu, Apr 16, 2009 at 17:09, Hyrum K. Wright
<hy...@mail.utexas.edu> wrote:
> Also keep in mind that some tools do a lot behind the scenes.  I
> believe Tortoise is fairly liberal with the information it fetches
> from the server, which may generate more distinct HTTP requests.
>
> I'm actually surprised that the command line client dominates as much
> as it does.  That may be a reflection of googlecode's target audience,
> though.
>
> On Apr 16, 2009, at 10:02 AM, C. Michael Pilato wrote:
>
>> Who knows why one client is used more than another, really?  Ben's
>> reporting
>> numbers which are, I'm guessing, based solely on counts of HTTP
>> requests.
>>
>> For example, could it be that SvnKit is what finds the greatest use in
>> continuous build tools?  If a continuous build tool is routinely
>> polling a
>> repository for changes, it could rack up more hits than the more
>> intentional
>> usage of TortoiseSVN.
>>
>>
>> Lieven Govaerts wrote:
>>> Interesting! I expected TortoiseSVN being used much more than SvnKit,
>>> but that's apparently only in corporate environments.
>>> Or is there an extraordinary high amount of Java projects hosted on
>>> googlecode?
>>>
>>> Lieven
>>>
>>> On Thu, Apr 16, 2009 at 4:44 PM, Ben Collins-Sussman
>>> <su...@red-bean.com> wrote:
>>>> googlecode gets a huge amount of svn traffic.  I was looking at log
>>>> statistics, specifically at aggregated user-agent: strings.  Here's
>>>> the relative popularity of different svn clients:
>>>>
>>>>
>>>> cmdline  65%
>>>> SvnKit   15%
>>>> Tortoise  8%
>>>> SharpSvn  1%
>>>> other    11%  (too many useragent micro-variants to count)
>>>>
>>>> cmdline breakdown:
>>>>
>>>> 1.6.x   5%
>>>> 1.5.x   50%
>>>> 1.4.x   30%
>>>> older   15%
>>>>
>>>> ------------------------------------------------------
>>>> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748379
>>>>
>>>
>>> ------------------------------------------------------
>>> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748459
>>
>>
>> --
>> C. Michael Pilato <cm...@collab.net>
>> CollabNet   <>   www.collab.net   <>   Distributed Development On
>> Demand
>>
>> ------------------------------------------------------
>> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748545
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748627
>

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748772


Re: interesting statistics

Posted by Karl Fogel <kf...@red-bean.com>.
"Hyrum K. Wright" <hy...@mail.utexas.edu> writes:
> I'm actually surprised that the command line client dominates as much  
> as it does.  That may be a reflection of googlecode's target audience,  
> though.

I don't believe that statistic, as a description of user experience.

Anyone writing a batch file (say, to regularly pull and build a branch)
is usually going to use the command-line client, for example.  So I
doubt that those user-agent strings reflect what humans are doing
interactively.

-K

> On Apr 16, 2009, at 10:02 AM, C. Michael Pilato wrote:
>
>> Who knows why one client is used more than another, really?  Ben's  
>> reporting
>> numbers which are, I'm guessing, based solely on counts of HTTP  
>> requests.
>>
>> For example, could it be that SvnKit is what finds the greatest use in
>> continuous build tools?  If a continuous build tool is routinely  
>> polling a
>> repository for changes, it could rack up more hits than the more  
>> intentional
>> usage of TortoiseSVN.
>>
>>
>> Lieven Govaerts wrote:
>>> Interesting! I expected TortoiseSVN being used much more than SvnKit,
>>> but that's apparently only in corporate environments.
>>> Or is there an extraordinary high amount of Java projects hosted on  
>>> googlecode?
>>>
>>> Lieven
>>>
>>> On Thu, Apr 16, 2009 at 4:44 PM, Ben Collins-Sussman
>>> <su...@red-bean.com> wrote:
>>>> googlecode gets a huge amount of svn traffic.  I was looking at log
>>>> statistics, specifically at aggregated user-agent: strings.  Here's
>>>> the relative popularity of different svn clients:
>>>>
>>>>
>>>> cmdline  65%
>>>> SvnKit   15%
>>>> Tortoise  8%
>>>> SharpSvn  1%
>>>> other    11%  (too many useragent micro-variants to count)
>>>>
>>>> cmdline breakdown:
>>>>
>>>> 1.6.x   5%
>>>> 1.5.x   50%
>>>> 1.4.x   30%
>>>> older   15%
>>>>
>>>> ------------------------------------------------------
>>>> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748379
>>>>
>>>
>>> ------------------------------------------------------
>>> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748459
>>
>>
>> -- 
>> C. Michael Pilato <cm...@collab.net>
>> CollabNet   <>   www.collab.net   <>   Distributed Development On  
>> Demand
>>
>> ------------------------------------------------------
>> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748545
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748627

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1752609

Re: interesting statistics

Posted by "Hyrum K. Wright" <hy...@mail.utexas.edu>.
Also keep in mind that some tools do a lot behind the scenes.  I  
believe Tortoise is fairly liberal with the information it fetches  
from the server, which may generate more distinct HTTP requests.

I'm actually surprised that the command line client dominates as much  
as it does.  That may be a reflection of googlecode's target audience,  
though.

On Apr 16, 2009, at 10:02 AM, C. Michael Pilato wrote:

> Who knows why one client is used more than another, really?  Ben's  
> reporting
> numbers which are, I'm guessing, based solely on counts of HTTP  
> requests.
>
> For example, could it be that SvnKit is what finds the greatest use in
> continuous build tools?  If a continuous build tool is routinely  
> polling a
> repository for changes, it could rack up more hits than the more  
> intentional
> usage of TortoiseSVN.
>
>
> Lieven Govaerts wrote:
>> Interesting! I expected TortoiseSVN being used much more than SvnKit,
>> but that's apparently only in corporate environments.
>> Or is there an extraordinary high amount of Java projects hosted on  
>> googlecode?
>>
>> Lieven
>>
>> On Thu, Apr 16, 2009 at 4:44 PM, Ben Collins-Sussman
>> <su...@red-bean.com> wrote:
>>> googlecode gets a huge amount of svn traffic.  I was looking at log
>>> statistics, specifically at aggregated user-agent: strings.  Here's
>>> the relative popularity of different svn clients:
>>>
>>>
>>> cmdline  65%
>>> SvnKit   15%
>>> Tortoise  8%
>>> SharpSvn  1%
>>> other    11%  (too many useragent micro-variants to count)
>>>
>>> cmdline breakdown:
>>>
>>> 1.6.x   5%
>>> 1.5.x   50%
>>> 1.4.x   30%
>>> older   15%
>>>
>>> ------------------------------------------------------
>>> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748379
>>>
>>
>> ------------------------------------------------------
>> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748459
>
>
> -- 
> C. Michael Pilato <cm...@collab.net>
> CollabNet   <>   www.collab.net   <>   Distributed Development On  
> Demand
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748545

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748627

Re: interesting statistics

Posted by "C. Michael Pilato" <cm...@collab.net>.
Who knows why one client is used more than another, really?  Ben's reporting
numbers which are, I'm guessing, based solely on counts of HTTP requests.

For example, could it be that SvnKit is what finds the greatest use in
continuous build tools?  If a continuous build tool is routinely polling a
repository for changes, it could rack up more hits than the more intentional
usage of TortoiseSVN.


Lieven Govaerts wrote:
> Interesting! I expected TortoiseSVN being used much more than SvnKit,
> but that's apparently only in corporate environments.
> Or is there an extraordinary high amount of Java projects hosted on googlecode?
> 
> Lieven
> 
> On Thu, Apr 16, 2009 at 4:44 PM, Ben Collins-Sussman
> <su...@red-bean.com> wrote:
>> googlecode gets a huge amount of svn traffic.  I was looking at log
>> statistics, specifically at aggregated user-agent: strings.  Here's
>> the relative popularity of different svn clients:
>>
>>
>> cmdline  65%
>> SvnKit   15%
>> Tortoise  8%
>> SharpSvn  1%
>> other    11%  (too many useragent micro-variants to count)
>>
>> cmdline breakdown:
>>
>> 1.6.x   5%
>> 1.5.x   50%
>> 1.4.x   30%
>> older   15%
>>
>> ------------------------------------------------------
>> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748379
>>
> 
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748459


-- 
C. Michael Pilato <cm...@collab.net>
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748545

Re: interesting statistics

Posted by Mark Phippard <ma...@gmail.com>.
SVNKit is used by a lot of Continuous Integration tools and these
constantly poll repositories for changes.  That will greatly skew
statistics.


On Thu, Apr 16, 2009 at 10:55 AM, Lieven Govaerts <sv...@mobsol.be> wrote:
> Interesting! I expected TortoiseSVN being used much more than SvnKit,
> but that's apparently only in corporate environments.
> Or is there an extraordinary high amount of Java projects hosted on googlecode?
>
> Lieven
>
> On Thu, Apr 16, 2009 at 4:44 PM, Ben Collins-Sussman
> <su...@red-bean.com> wrote:
>> googlecode gets a huge amount of svn traffic.  I was looking at log
>> statistics, specifically at aggregated user-agent: strings.  Here's
>> the relative popularity of different svn clients:
>>
>>
>> cmdline  65%
>> SvnKit   15%
>> Tortoise  8%
>> SharpSvn  1%
>> other    11%  (too many useragent micro-variants to count)
>>
>> cmdline breakdown:
>>
>> 1.6.x   5%
>> 1.5.x   50%
>> 1.4.x   30%
>> older   15%
>>
>> ------------------------------------------------------
>> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748379
>>
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748459
>



-- 
Thanks

Mark Phippard
http://markphip.blogspot.com/

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748533


Re: interesting statistics

Posted by Lieven Govaerts <sv...@mobsol.be>.
Interesting! I expected TortoiseSVN being used much more than SvnKit,
but that's apparently only in corporate environments.
Or is there an extraordinary high amount of Java projects hosted on googlecode?

Lieven

On Thu, Apr 16, 2009 at 4:44 PM, Ben Collins-Sussman
<su...@red-bean.com> wrote:
> googlecode gets a huge amount of svn traffic.  I was looking at log
> statistics, specifically at aggregated user-agent: strings.  Here's
> the relative popularity of different svn clients:
>
>
> cmdline  65%
> SvnKit   15%
> Tortoise  8%
> SharpSvn  1%
> other    11%  (too many useragent micro-variants to count)
>
> cmdline breakdown:
>
> 1.6.x   5%
> 1.5.x   50%
> 1.4.x   30%
> older   15%
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748379
>

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1748459


Re: interesting statistics

Posted by Stefan Küng <to...@gmail.com>.
Unfortunately, I don't think there is a way to distinguish svn 1.4.x
from tsvn 1.4.x (or other svn clients linked to 1.4.x or earlier).


Ben Collins-Sussman wrote:
> I think you're right -- I think a bunch of TSVN traffic is buried in
> that 'cmdline' statistic.  Is there any way to distinguish svn 1.4.x
> from tsvn 1.4.x?
> 
> 
> On Sat, Apr 18, 2009 at 12:19 PM, Stefan Küng <to...@gmail.com> wrote:
>> Ben Collins-Sussman wrote:
>>> googlecode gets a huge amount of svn traffic.  I was looking at log
>>> statistics, specifically at aggregated user-agent: strings.  Here's
>>> the relative popularity of different svn clients:
>>>
>>>
>>> cmdline  65%
>>> SvnKit   15%
>>> Tortoise  8%
>>> SharpSvn  1%
>>> other    11%  (too many useragent micro-variants to count)
>>>
>>> cmdline breakdown:
>>>
>>> 1.6.x   5%
>>> 1.5.x   50%
>>> 1.4.x   30%
>>> older   15%
>> Ben, since 45% of clients are older than 1.5.x, are the stats for the
>> different clients over all versions or just the >= 1.5.x clients?
>> It's because svn only offers customizing the user agent string since
>> 1.5.0 - for earlier versions all clients would show up as 'cmdline' (not
>> sure about SvnKit - I think they might have used their own string even
>> before 1.5.0).
>>
>> If it's for all versions, then the client stats are a little bit skewed...
>>
>> Stefan
>>
>> --
>>       ___
>>  oo  // \\      "De Chelonian Mobile"
>>  (_,\/ \_/ \     TortoiseSVN
>>   \ \_/_\_/>    The coolest Interface to (Sub)Version Control
>>   /_/   \_\     http://tortoisesvn.net
>>
>>


-- 
       ___
  oo  // \\      "De Chelonian Mobile"
 (_,\/ \_/ \     TortoiseSVN
   \ \_/_\_/>    The coolest Interface to (Sub)Version Control
   /_/   \_\     http://tortoisesvn.net

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1818376

Re: interesting statistics

Posted by Ben Collins-Sussman <su...@red-bean.com>.
I think you're right -- I think a bunch of TSVN traffic is buried in
that 'cmdline' statistic.  Is there any way to distinguish svn 1.4.x
from tsvn 1.4.x?


On Sat, Apr 18, 2009 at 12:19 PM, Stefan Küng <to...@gmail.com> wrote:
> Ben Collins-Sussman wrote:
>> googlecode gets a huge amount of svn traffic.  I was looking at log
>> statistics, specifically at aggregated user-agent: strings.  Here's
>> the relative popularity of different svn clients:
>>
>>
>> cmdline  65%
>> SvnKit   15%
>> Tortoise  8%
>> SharpSvn  1%
>> other    11%  (too many useragent micro-variants to count)
>>
>> cmdline breakdown:
>>
>> 1.6.x   5%
>> 1.5.x   50%
>> 1.4.x   30%
>> older   15%
>
> Ben, since 45% of clients are older than 1.5.x, are the stats for the
> different clients over all versions or just the >= 1.5.x clients?
> It's because svn only offers customizing the user agent string since
> 1.5.0 - for earlier versions all clients would show up as 'cmdline' (not
> sure about SvnKit - I think they might have used their own string even
> before 1.5.0).
>
> If it's for all versions, then the client stats are a little bit skewed...
>
> Stefan
>
> --
>       ___
>  oo  // \\      "De Chelonian Mobile"
>  (_,\/ \_/ \     TortoiseSVN
>   \ \_/_\_/>    The coolest Interface to (Sub)Version Control
>   /_/   \_\     http://tortoisesvn.net
>
>

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1816065


Re: interesting statistics

Posted by Stefan Küng <to...@gmail.com>.
Ben Collins-Sussman wrote:
> googlecode gets a huge amount of svn traffic.  I was looking at log
> statistics, specifically at aggregated user-agent: strings.  Here's
> the relative popularity of different svn clients:
> 
> 
> cmdline  65%
> SvnKit   15%
> Tortoise  8%
> SharpSvn  1%
> other    11%  (too many useragent micro-variants to count)
> 
> cmdline breakdown:
> 
> 1.6.x   5%
> 1.5.x   50%
> 1.4.x   30%
> older   15%

Ben, since 45% of clients are older than 1.5.x, are the stats for the
different clients over all versions or just the >= 1.5.x clients?
It's because svn only offers customizing the user agent string since
1.5.0 - for earlier versions all clients would show up as 'cmdline' (not
sure about SvnKit - I think they might have used their own string even
before 1.5.0).

If it's for all versions, then the client stats are a little bit skewed...

Stefan

-- 
       ___
  oo  // \\      "De Chelonian Mobile"
 (_,\/ \_/ \     TortoiseSVN
   \ \_/_\_/>    The coolest Interface to (Sub)Version Control
   /_/   \_\     http://tortoisesvn.net

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1793052