comp.lang.ada
 help / color / mirror / Atom feed
From: gabrielb@rational.com (Gabriel Bereny)
Subject: Re: IDE/Rational Apex experiences
Date: 1997/11/11
Date: 1997-11-11T00:00:00+00:00	[thread overview]
Message-ID: <64a8tp$lro$1@rational2.rational.com> (raw)
In-Reply-To: 6429gb$5qh@dfw-ixnews12.ix.netcom.com


bklungle (bklungle@ix.netcom.com) wrote:
: At a certain large AeroSpace company, use of Rational Apex was mandated by
: management. There are 27 programmers trying to develop a very large ground
: based processing system using C and C++. When 4 users log onto the Rational
: Apex system, the response time on the network increases to around 20
: minutes. The system HAS been optimized about as much as the Gurus know how.
: At this time, there is a pending uprising within the working staff to
: eliminate it. Instead of a productivity improvement, it has brought
: development to a standstill. Almost everyone would rather go back to SCCS
: than continue with Apex.
: One users experience over the past 11 months.

Mr. Klungle does not specify anything regarding the hardware on which
his company is running Apex, nor about the network configuration, nor
about what sorts of operations are being performed on the system.
I'm not doubting the veracity of his statement that things are slow
or that the system has been optimized "as much as the gurus know how".
But still we are left with questions about the precise nature of the
system abd the operations.

If Rational Apex is installed on a single IPC with 64 MB of RAM and
4 users access Apex using PCs running some xterm emulation software
as displays and rlogins to the IPC to run Apex, I have no doubt that
response will be miserable. Note that this still fits in with the
information that Mr. Klungle supplies. The system described is not
what one would recommend, but it may well be "optimized" as much as
it can be. 

We in Rational technical support have run six, seven, and more
sessions of Apex on a single server (first a SPARCstation 10, now
an UltraSPARC 2) using other SPARCs as displays many times over the 
past few years and have never seen "response time on the network 
increase to around 20 minutes" (whatever that statement may mean).

If Mr. Klungle claims that his company has a proper hardware and
network setup and that response time for an arbitrary operation
(again, there's no clue about what operations he's discussing) is
20 minutes, then I would have to doubt that the system is
optimized or that the operation to which he is referring is
"normal". "Not normal" might include checking back into a library
a 400,000 line file in which every other line has been changed.

I agree that Apex is large and slower than something like SCCS. 
On the other hand, Apex does a heck of a lot more than SCCS, than
vi, than ...  There's no free lunch. Increased functionality
requires larger amounts of space and large executables require
more resources to run. This is certainly not a quality unique to
Apex. If the current resources are strained even though they have
been optimized, then more resources are needed. This is the way
of the world in the computer business. Rational didn't make it
this way but neither can Rational change this immutable fact.

If the management of "large AeroSpace company" bought Rational Apex 
strictly as a replacement for SCCS, I'd be surprised. If SCCS is
all that is needed, the working staff has a legitimate gripe about
the response time. But I think that time griping and waiting for
system response would be better spent presenting a case to the
management that if they expect software tools to increase productivity,
then they must invest in some hardware as well.



Gabriel Bereny
Rational Software Corporation
Technical Support




  parent reply	other threads:[~1997-11-11  0:00 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
1997-11-07  0:00 IDE/Rational Apex experiences Damian O'Neill
1997-11-08  0:00 ` bklungle
1997-11-08  0:00   ` Robert S. White
1997-11-10  0:00     ` bklungle
1997-11-11  0:00       ` Robert S. White
1997-11-10  0:00   ` Robert B. Love 
1997-11-10  0:00     ` bklungle
1997-11-10  0:00       ` David  Weller
1997-11-11  0:00   ` Gabriel Bereny [this message]
1997-11-12  0:00     ` bklungle
1997-11-12  0:00 ` KJPrice
1997-11-14  0:00   ` Don Harrison
1997-12-02  0:00     ` TB
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox