[p4] How does Bitkeeper compare to Perforce?

Tue Jul 3 10:31:44 PDT 2007

I'm no expert, so I can't comment on a lot of the claims on the site,
but I would like to add my 2c where I can.     I totally agree with
Jeff:

>Which brings to what I consider an outright lie:
>>The Perforce ... database ... get[s] corrupted frequently
>In 7 years of continuous operation we have never seen the database get
corrupted.  AFAIK, this is a common experience. 

Same here - 5 years, not one byte out of place.  I imagine a very badly
maintained Perforce system could become corrupted, but I don't think any
SCM could function usefully under those circumstances.

Oh, and as for the comment about maintenance - we have 3 Perforce admins
here.  All of us have other roles, and we spend less than an hour a week
between us looking after it - some weeks we don't need to touch it.  It
took less than a week to install the first version and get it working
(with backups etc) and the last upgrade took less than 3 man days to
plan, rehearse and execute.  I don't know if Bitkeeper is even more
simple to look after, but even if it took half as much maintenance,
you're looking at a saving of 20-30 minutes a week!

Regards, Nic

-----Original Message-----
From: perforce-user-bounces at perforce.com
[mailto:perforce-user-bounces at perforce.com] On Behalf Of Smith, Jeff
Sent: 03 July 2007 15:05
To: Jeremy.Chatelaine; perforce-user at perforce.com
Subject: Re: [p4] How does Bitkeeper compare to Perforce?

No doubt others will chime in.  This is standard marketing FUD.  Some of
it is simply misleading while others are outright false.  A few
examples:

> The database is a single point of failure; if it gets corrupted, your
source management system does not work.
This is an overstatement.  It's kinda like saying that one of the flaws
in an automobile is that there is only one steering wheel.  If anything
happens to the driver while you are driving, you have serious problems.
While true, it doesn't happen often enough to be of any concern.  And
with Perforce, database corruption is an extremely rare occurrence.

Which brings to what I consider an outright lie:
>The Perforce ... database ... get[s] corrupted frequently
In 7 years of continuous operation we have never seen the database get
corrupted.  AFAIK, this is a common experience.

>Atomic ChangeSets: No
Maybe if you use a specialized definition of "Atomic ChangeSets" but
Perforce changelists meet any normal SCM definition.

>Merging in Perforce is primitive at best. 
>Graphical checkin tool: Weak
Maybe in the past but current version of P4Win/P4Visual are very good in
this respect.

>Upgrades are not reversible and lock the system for hours.
Maybe on really large DBs but our upgrades are measured in minutes.  And
this only for major release upgrades.  Maintenance releases don't
upgrade the DB schema so this is not an issue for any size installation.

>Perforce has an integrity checker but it is only run if you ask for it,
i.e., the default is to just hope that the data is correct.
This is true only if you don't read the Administrator documentation and
don't follow best practices: run verify regularly usually with a
scheduled job.  The implication here is that Bitkeeper requires no
similar administrative care and feeding.  While that may true, I doubt
it.  I have yet to meet a truly functional SCM system which is able to
run without and admin to look after it.

Which bring us to:
>Minimal Administration: Headcount can be used for doing development
rather than upkeep of the SCM system.
This one should scare you.  If software development is important to your
business then your company should acknowledge that.  We're talking about
managing the intellectual property of your company.  Trusting a
hands-off system with that data and no in-house specialist who
understands that system and looks after it daily is foolhardy at
best...disastrous at worst.  This is thrown in to convince management
types to want the system with the promise of "low-cost SCM" which is
too-often seen as simply a necessary evil.

>Perforce is marketed as the fast SCM system but it chooses speed over
correctness.
This is just plain stupid.  They are trying to make you think Perforce
is unreliable and that you should fear data corruption as a common
occurrence...simply not true.

>Perforce uses the RCS file format with all of the problems that
entails.
More FUD.  Only partially true.  Most of the problems with the RCS file
format is that the metadata was kept there.  Perforce keeps that in
their DB.  

>The database can use a dramatic amount of disk space. 
Only because the tracking information is kept there.  If not there then
where?  What is important is not the size of one piece of the system.
Assuming that BitKeeper is keeping just as much info (or more since they
claim to track comments per file) then I'm not sure how their overall
repository size could be less.

>Dynamic Licensing: No
By their description Perforce has the same functionality: Provides
developers the flexibility of checking in from any host or domain.  If
they mean floating license then Perforce doesn't offer that model.  I
note that they don't indicate pricing on the website.  I suspect your
total license cost for BitKeeper will be higher that for Perforce.

>Accurate recording of all history: No
Another lie.  Perforce does accurately record all history.  Another
attempt to scare you into believing that Perforce will loose or corrupt
data.

Summary:
Perforce does have its flaws and weaknesses (lack of directory
revisioning and reflecting branches as a first-class object of the
system) but reliability is not one of them.  I know nothing about
BitKeeper but if this is what they feel they need to do to sell their
product, I'd be wary.

Jeff

-----Original Message-----
From: perforce-user-bounces at perforce.com
[mailto:perforce-user-bounces at perforce.com] On Behalf Of
Jeremy.Chatelaine
Sent: Tuesday, July 03, 2007 8:52 AM
To: perforce-user at perforce.com
Subject: [p4] How does Bitkeeper compare to Perforce?

Hi all,

I'm a happy Perforce user but recently the company I work for ran some
source control evaluation.

One person pointed me to this page:
http://www.bitkeeper.com/Comparisons.Perforce.html

I was surprised to some points in the document to say the least and I
was wondering if anyone here has anything to say about BitKeeper
compared to Perforce.

I looked on Perforce website for a comparison document but could not
find any. Posts on the mailing list are ether old or not greatly
addressing this question.

So, can anyone let me know more information about BitKeeper vs Perforce?

I would be happy to hear about some real production experiences too.

Thanks a lot in advance,

Jeremy
http://jeremy.chatelaine.name

_______________________________________________
perforce-user mailing list  -  perforce-user at perforce.com
http://maillist.perforce.com/mailman/listinfo/perforce-user

_______________________________________________
perforce-user mailing list  -  perforce-user at perforce.com
http://maillist.perforce.com/mailman/listinfo/perforce-user