On Sat, Jul 22, 2006 at 11:36:42AM -0700, Stewart Stremler wrote:
> begin quoting James G. Sack (jim) as of Fri, Jul 21, 2006 at 07:05:45PM
> -0700:
> > Paul G. Allen wrote:
> > >..
> > > I use the free version of Perforce
> > >..
> >
> > Do you use their gui interface, or have you tried it?
> >
> > If so, I'm wondering how easy/useful it might be for users who are
> > unsophisticated in scm concepts.
>
> I think it takes a week to come up to speed, no matter the interface.
> Pick the GUI or the CLI depending on the user's preference and
> capabilities, not because one or the other is intrinsically "better".
>
> I found both the CLI and GUI tools to be useful, with the CLI being
> better for some tasks, and the GUI being better for others. However,
> in trying to get developers up and using a tool, the grasping of basic
> concepts and routine is the hard-part; basic use of the tool doesn't
> seem to matter on the tool.
>
> Although, there's a lot less whining about how "hard" a command-line
> is with a GUI tool, actual proficiency seems to take about the same
> amount of time (in my experience). Some folks will be faster, some
> slower, and no matter what the features are of the tool or interface,
> there will be someone who will back out a week's changes 'cuz they
> didn't bother to lisen.
>
<sigh> I've been doing this for a living for over 15 years. Here are my
observations.
- Developers who have never used an SCM tool (especially young ones)
resist it as a waste of time ("slow us down"). Developers with
experience with it are grateful to have it added to the project.
Likewise managers. When someone says "waste of time" or "slow us down" I
know he's not really a pro yet.
- Old timers and embedded programmers are the most eager to accept SCM
disciplines, Micro$oft-only types are the biggest whiners, especially
about integration into their precious IDEs. I've watched supposedly
intelligent designers waste 90 minutes of a meeting doing meticulous
calculations to show that two keystrokes per compile is an unacceptable
burden on the company's bottom line. God, I wish I'd video taped it!
- Nobody in development knows or understands SCM. It isn't taught in
schools. The academics are just discovering it. The few books are either
arcane or absurdly simplistic. The issues go way deeper than saving and
recovering code. For example, in a task-driven model, issue/bug tracking
is an integral part of the SCM imperatives (traceability, repeatability,
accountability).
Properly done, SCM should stay out of development's way as much as
possible and yet solve problems before they happen. My boss (he's one of
the best) says we should "be invisible and bullet-proof." Sounds like
superpowers to me.
I am constantly learning, which keeps me happy. But I don't think I
would have any patience at this stage in my life with working at a shop
in which SCM gets thrown overboard at every schedule crunch because "it
just gets in the way." Because next month's unexpected and insoluble
problem frequently points back to this month's SCM shortcut.
Oh, we use Perforce. It's pretty good. p4v is superior to p4win, and
yes, the command line is all you really need if you think that way.
--
Lan Barnes
Linux Guy, SCM Specialist
Tcl/Tk Enthusiast
[A]ll our science, measured against reality, is primitive and childlike
-- and yet it is the most precious thing we have.
- Albert Einstein
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
