Path: utzoo!attcan!uunet!ogicse!dali!uakari.primate.wisc.edu!zaphod.mps.ohio-state.edu!usc!apple!voder!pyramid!leadsv!cberg
From: cberg@leadsv.UUCP (Charles R Berg)
Newsgroups: comp.software-eng
Subject: Re: Contemplating Purchase of "Excelerator" CASE Tool
Message-ID: <11422@leadsv.UUCP>
Date: 1 Jun 90 16:53:18 GMT
Reply-To: cberg@leadsv.LEADS.LMSC.COM.UUCP (Charles R Berg)
Organization: LMSC-LEADS, Sunnyvale, Ca.
Lines: 69

In article <11394@leadsv.UUCP> cberg@leadsv.LEADS.LMSC.COM.UUCP (I) wrote:
>In some article, somebody wrote:
>>Can you compare Excelerator with any other CASE tools for the same job?
>
>I used Excelerator for about 18 months (3 years ago), and am now using StP
>for about 2 years.
>
>First, some philosophy.  (philosophy deleted)

One other significant 'philosophy' difference.  Excelerator interacts with the
database on-line during the diagram editting process.  With StP, you edit the
diagrams 'off-line' and in a separate step, 'compile' them into the database.

This results in some interesting trade-offs.  With Excelerator, if you can't
remember the name of the module you want to reference?  A few keystrokes and
mouse-clicks and you have a list on the screen from which to select.  It
automatically restricts the list to only those objects (by type) that are
valid for the current field, and names can be qualified by wild-card.  With
StP, the nearest equivalent is the Database Browser.  This provides a
scrollable list of all objects in the database.  If there are any mechanisms
to qualify this list by object type or wild-card name selection, I'm not aware
of any.

Another trade-off.  With StP, when you create a diagram, you frequently
reference existing objects in the database.  Data structure diagrams reference
existing data structures, structure charts reference existing modules, etc.
These references can be annotated with all kinds of free form and strucured
information, but only the annotation on the 'defining' object goes into the
database.  Virtually the reverse is true with module parameters on a structure
chart.  If the module is defined on a separate diagram, with the parameters
entering from 'off-page', these parameters are NOT entered into the database.
But on other diagrams, where references to this module are made, the parameters
ARE entered into the database.  And, there are no checks to ensure that all
references to a given module use the same parameters.  (But you could easily
write your own DPS script to verify this.)

With Excelerator, during the editting process, any reference to a defined
object will automatically provide the existing database entries for that
object.  If you change the (Excelerator equivalent of) annotation, you are
changing the one and only one copy of that information that exists.

These examples imply (to me anyway) that the Excelerator approach is better.
However, there is a big price you pay for it.  Excelerator requires that each
user 'check-out' that portion of the database that they want to work with.
Then they can do their job, and check it back in.  This is more than just
checking out one diagram at a time.  It means checking out all the diagrams and
objects for a major portion of the system, since thats the only way that all
the objects you might want to reference will be in the local database.  And
there is still the opportunity for conflict when two or more people copy
out information for reference purpose, and then want to check it all back in.
There are procedures for taking care of all this, but it really requires some
effort on the part of a database administrator person.

With StP, there is a single database, (usually on a network), that is
dynamically locked and unlocked as each user interacts with it.  In addition,
individual diagrams are dynamically locked and unlocked as they are 'checked-
out' for editing.

So, once again, which is better depends on individual requirements.  My first
project only involved three users of the Excelerator database.  We didn't
bother checking out individual copies of the database for private use.  We
just hollered around the room as we worked.  With StP, we have 12 or so users,
so the informal approach would have been very difficult to rely on.

Once again, I hope this information helps.

Chuck Berg

PS: Same caveats and disclaimers as before.