You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@subversion.apache.org by Kenneth Porter <sh...@sewingwitch.com> on 2006/08/21 22:26:36 UTC
Perforce versus Subversion
A coworker pointed out this document on the Perforce website:
<http://www.perforce.com/perforce/comparisons/perforce_subversion.pdf>
Except for the benchmarks reported, Perforce just claims to do things
differently (but worded in a way to make it sound like they're better, if
you know nothing about Subversion or FOSS).
Has anyone responded to this?
My biggest concern with a commercial product is a proprietary closed data
format. The Subversion dump format gives me a lot of peace of mind on this
issue. I haven't seen any indication that Perforce offers something
comparable.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Perforce versus Subversion
Posted by Justin Erenkrantz <ju...@erenkrantz.com>.
On 8/22/06, Arild Fines <ar...@broadpark.no> wrote:
> >> Subversion 1.3
>
> >> * Running a full branch on same files 4m 3.119s
>
> Eh. How did they manage to spend 4 minutes doing a branch with Subversion?
The best that we can come up with is that they did a WC->WC 'svn cp' -
which isn't how we would recommend creating a branch - it should be
URL->URL.
Coming from a P4 perspective, I can see why they would never ever
think of doing a branch ops on URLs. -- justin
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
RE: Perforce versus Subversion
Posted by Arild Fines <ar...@broadpark.no>.
> -----Original Message-----
> From: Kenneth Porter [mailto:shiva@sewingwitch.com]
> Sent: Tuesday, August 22, 2006 12:27 AM
> To: users@subversion.tigris.org
> Subject: Perforce versus Subversion
>
> A coworker pointed out this document on the Perforce website:
>
> <http://www.perforce.com/perforce/comparisons/perforce_subversion.pdf>
Re: Perforce versus XYZ
Posted by Kenneth Porter <sh...@sewingwitch.com>.
Another article with links to pages comparing Subversion to competitors:
<http://www.szabgab.com/subversion_vs_xyz.html>
(This also includes links back into the archives for this mailing list.)
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Perforce versus Subversion
Posted by Justin Erenkrantz <ju...@erenkrantz.com>.
On 8/21/06, Kenneth Porter <sh...@sewingwitch.com> wrote:
> Has anyone responded to this?
At least Perforce is admitting that we exist now. ;-) -- justin
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Perforce versus Subversion
Posted by Sean Kelley <se...@gmail.com>.
On 8/23/06, Garrett Rooney <ro...@electricjellyfish.net> wrote:
> On 8/21/06, Kenneth Porter <sh...@sewingwitch.com> wrote:
>
> > > Scalability and performance: Limited due to file binary diff
> > > versioning process.
> >
> > Unknown. [Seeking developer input.]
> >
>
> Ironically, whenever I've talked to Perforce users their big complaint
> seems to be scalability, p4 really falls down with many clients that
> have large working copies, due to the amount of state it needs to keep
> in memory on the server.
>
> As for binary diffing being a performance problem for svn, I think
> they're kind of stretching there...
I evaluated Perforce, having used in the past both StarTeam and now
Subversion. I found Perforce to be very similar to StarTeam in terms
of its heavy use of the network connection and complex GUI client.
LIke StarTeam it requires you to first "Edit" or "Checkout" the file.
It keeps two different revisions (?) with a file, the depot revision
and the changeset revision perhaps. It is very confusing. I am used
to working with StarTeam and now Subversion.
Sean
>
> -garrett
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
> For additional commands, e-mail: users-help@subversion.tigris.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Perforce versus Subversion
Posted by Garrett Rooney <ro...@electricjellyfish.net>.
On 8/21/06, Kenneth Porter <sh...@sewingwitch.com> wrote:
> > Scalability and performance: Limited due to file binary diff
> > versioning process.
>
> Unknown. [Seeking developer input.]
>
Ironically, whenever I've talked to Perforce users their big complaint
seems to be scalability, p4 really falls down with many clients that
have large working copies, due to the amount of state it needs to keep
in memory on the server.
As for binary diffing being a performance problem for svn, I think
they're kind of stretching there...
-garrett
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Perforce versus Subversion
Posted by eg <eg...@gmail.com>.
Kenneth Porter wrote:
>
>> Scalability and performance: Limited due to file binary diff
>> versioning process.
>
> Unknown. [Seeking developer input.]
>
I noticed from the SVNSummit agenda notes, that Dan Berlin is looking at
alternate FSFS backends. Perhaps something based on mercurials revlog
format.
You can read a paper about the revlog format at:
http://linux.inet.hr/ols2006_towards_a_better_scm_revlog_and_mercurial.html
Very cool work!
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Perforce versus Subversion
Posted by Garrett Rooney <ro...@electricjellyfish.net>.
On 8/22/06, Kenneth Porter <sh...@sewingwitch.com> wrote:
> --On Monday, August 21, 2006 11:06 PM -0700 Karl Fogel <kf...@google.com>
> wrote:
>
> > Kenneth Porter <sh...@sewingwitch.com> writes:
> >> True. Subversion's merge tracking is in the conceptual planning stage:
> >>
> >> <http://subversion.tigris.org/merge-tracking/>
> >
> > It's gotten a bit farther than that now -- people are actively working
> > on it, especially Daniel Berlin. See
> >
> > http://svn.collab.net/repos/svn/branches/merge-tracking/
>
> I poked around in the notes and doc directories but don't see anything
> obvious describing the merge-tracking architecture. Is there any equivalent
> branch for the Book that would let me see what this is going to look like,
> both conceptually and operationally? (The /branches directory for the Book
> is empty.)
Nothing gets documented in the book until it's actually done. The
merge tracking docs live on the web site at the moment:
http://subversion.tigris.org/merge-tracking/
Although keep in mind that nothing in there is final.
-garrett
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Perforce versus Subversion
Posted by Kenneth Porter <sh...@sewingwitch.com>.
--On Monday, August 21, 2006 11:06 PM -0700 Karl Fogel <kf...@google.com>
wrote:
> Kenneth Porter <sh...@sewingwitch.com> writes:
>> True. Subversion's merge tracking is in the conceptual planning stage:
>>
>> <http://subversion.tigris.org/merge-tracking/>
>
> It's gotten a bit farther than that now -- people are actively working
> on it, especially Daniel Berlin. See
>
> http://svn.collab.net/repos/svn/branches/merge-tracking/
I poked around in the notes and doc directories but don't see anything
obvious describing the merge-tracking architecture. Is there any equivalent
branch for the Book that would let me see what this is going to look like,
both conceptually and operationally? (The /branches directory for the Book
is empty.)
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Perforce versus Subversion
Posted by Karl Fogel <kf...@google.com>.
Kenneth Porter <sh...@sewingwitch.com> writes:
> True. Subversion's merge tracking is in the conceptual planning stage:
>
> <http://subversion.tigris.org/merge-tracking/>
It's gotten a bit farther than that now -- people are actively working
on it, especially Daniel Berlin. See
http://svn.collab.net/repos/svn/branches/merge-tracking/
Best,
-Karl
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Perforce versus Subversion
Posted by Kenneth Porter <sh...@sewingwitch.com>.
Their document contains a table comparing aspects of Subversion to
Perforce. I compiled what I knew about Subversion in response:
> Branching and merging: Subversion offers branching, but requires
> manual tracking of merge history.
True. Subversion's merge tracking is in the conceptual planning stage:
<http://subversion.tigris.org/merge-tracking/>
> Platform dependencies: Compatibility issues exist with line ending of
> files in Unix and Windows environments.
False. Subversioned files have an optional line ending property that's used
to determine whether to convert the file to the native format on checkout:
<http://svnbook.red-bean.com/nightly/en/
svn.advanced.props.html#svn.advanced.props.special.eol-style>
> Distributed development: No performance solutions for remote
> development are available.
False. Subversion was designed for efficient use over a wide-area network,
and minimizes traffic to the repository. Most common queries can be
performed without any network access at all.
> Scalability and performance: Limited due to file binary diff
> versioning process.
Unknown. [Seeking developer input.]
> Defect tracking: Not part of Subversion; need to use a separate defect
> tracking solution.
True. Multiple solutions are available. Trac is well-regarded:
<http://trac.edgewall.org/>
> Integration with related tools: Integrations are available via the
> open source community.
True. Integration with Windows Explorer is available through TortoiseSVN.
Integration with Visual Studio is available through AnkhSVN.
<http://tortoisesvn.tigris.org/>
<http://ankhsvn.tigris.org/>
> Support: Available at additional cost from a third party.
True. But note that the base cost is zero. So you only pay for support, not
the code. Plus, you get full source, in case you need a feature the
developer can't or won't supply.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Perforce versus Subversion
Posted by Jeremy Pereira <je...@jeremyp.net>.
From the PDF:
"Perforce is architected first and foremost for speed."
Were that anything more than just marketing droid speak, I would
chuck out Perforce on the basis of that sentence alone. I want my
version control system to be architected first and foremost for
keeping all of my code versions secure. Speed is a secondary
consideration.
On 21 Aug 2006, at 23:26, Kenneth Porter wrote:
> A coworker pointed out this document on the Perforce website:
>
> <http://www.perforce.com/perforce/comparisons/perforce_subversion.pdf>
>
> Except for the benchmarks reported, Perforce just claims to do
> things differently (but worded in a way to make it sound like
> they're better, if you know nothing about Subversion or FOSS).
>
> Has anyone responded to this?
>
> My biggest concern with a commercial product is a proprietary
> closed data format. The Subversion dump format gives me a lot of
> peace of mind on this issue. I haven't seen any indication that
> Perforce offers something comparable.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
> For additional commands, e-mail: users-help@subversion.tigris.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: Perforce versus Subversion
Posted by Nico Kadel-Garcia <nk...@comcast.net>.
Chris.Fouts@qimonda.com wrote:
> Subversion is FREE, Perforce is NOT for "commercial"
> use. End of evaluation IMHO, agin for commercial use.
> I'm not saying that makes Subversion "better", but
> it does the job adequately - except for cumbersome
> branching/merging.
>
> I've evaluated BOTH, and chose Subversion at work
> ONLY for that reason. If only for personal use, I'd
> probably use Perforce.
Perforce also used to have a huge bug with symlinks, which I've asked them
about since and they've never said is fixed. If you edit a symlink by
relinking it to a new target, then commit the change, then check out a fresh
copy of the committed directory, you wind up with the old symlink. The only
way to change a symlink was to actually delete it, then add a new one.
Subversion gets this right.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
RE: RE: Perforce versus Subversion
Posted by Kenneth Porter <sh...@sewingwitch.com>.
--On Tuesday, August 22, 2006 6:04 PM +0200 Chris.Fouts@qimonda.com wrote:
> Perforce - NOT free for commercial use, meaning support comes with
> the license fees
Agreed.
My confusion was over your insistence on a free license for commercial use.
I don't consider it inappropriate for someone to charge me for something
they've created. And if I'm making a profit, presumably I have money to
spend on tools. So the fact that software costs money shouldn't preclude it
from consideration, esp. in a commercial context.
While cheaper initial cost is better, it's not the only criteria I use in
selecting my tools. (The corporate-speak for this is "TCO".)
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
RE: RE: Perforce versus Subversion
Posted by Ch...@qimonda.com.
>-----Original Message-----
>From: Kenneth Porter [mailto:shiva@sewingwitch.com]
>Sent: Tuesday, August 22, 2006 11:58 AM
>To: users@subversion.tigris.org
>Subject: RE: Perforce versus Subversion
>
>--On Tuesday, August 22, 2006 3:25 PM +0200
>Chris.Fouts@qimonda.com wrote:
>
>> Subversion is FREE, Perforce is NOT for "commercial"
>> use. End of evaluation IMHO, agin for commercial use.
>
>Did you mean non-commercial?
>
>Big companies seeking SCM also want support. Subversion has
>both free support (this list) and paid support from companies
>such as CollabNet. If you're a corporate manager seeking the
>"warm fuzzies" of an invoice and paid training, you can go to
>CollabNet and get that.
>
>This is a Good Thing. It means you get all the benefits of
>FOSS we've come to love, plus all the benefits of traditional
>closed-source products.
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
>For additional commands, e-mail: users-help@subversion.tigris.org
>
Perforce - NOT free for commercial use, meaning support comes with
the license fees
Subversion - free for ANY use; paid support is optional
-chris
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
RE: Perforce versus Subversion
Posted by Kenneth Porter <sh...@sewingwitch.com>.
--On Tuesday, August 22, 2006 3:25 PM +0200 Chris.Fouts@qimonda.com wrote:
> Subversion is FREE, Perforce is NOT for "commercial"
> use. End of evaluation IMHO, agin for commercial use.
Did you mean non-commercial?
Big companies seeking SCM also want support. Subversion has both free
support (this list) and paid support from companies such as CollabNet. If
you're a corporate manager seeking the "warm fuzzies" of an invoice and
paid training, you can go to CollabNet and get that.
This is a Good Thing. It means you get all the benefits of FOSS we've come
to love, plus all the benefits of traditional closed-source products.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
RE: Perforce versus Subversion
Posted by Ch...@qimonda.com.
Subversion is FREE, Perforce is NOT for "commercial"
use. End of evaluation IMHO, agin for commercial use.
I'm not saying that makes Subversion "better", but
it does the job adequately - except for cumbersome
branching/merging.
I've evaluated BOTH, and chose Subversion at work
ONLY for that reason. If only for personal use, I'd
probably use Perforce.
Perforce supports a "configuration spec" idea, similar
to Clearcase (but not quite) though, which neither CVS
nor Subversion sadly do NOT.
-chris
>-----Original Message-----
>From: Kenneth Porter [mailto:shiva@sewingwitch.com]
>Sent: Monday, August 21, 2006 6:27 PM
>To: users@subversion.tigris.org
>Subject: Perforce versus Subversion
>
>A coworker pointed out this document on the Perforce website:
>
><http://www.perforce.com/perforce/comparisons/perforce_subversion.pdf>
>
>Except for the benchmarks reported, Perforce just claims to do
>things differently (but worded in a way to make it sound like
>they're better, if you know nothing about Subversion or FOSS).
>
>Has anyone responded to this?
>
>My biggest concern with a commercial product is a proprietary
>closed data format. The Subversion dump format gives me a lot
>of peace of mind on this issue. I haven't seen any indication
>that Perforce offers something comparable.
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
>For additional commands, e-mail: users-help@subversion.tigris.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org