Case tool

Daniel Lemire Daniel.Lemire at Videotron.ca
Fri Dec 1 23:52:57 UTC 2000


I tend to agree.

My general conclusion is that if you can afford a *very* powerful
machine, then Together is just fine. It means you must be able to do
away with 128 Megs without noticing it. It implies you have 512 Megs at
least. And you need one of the faster processor on the market. (Even a
single 800 Mhz processor might not be quite enough, but it will help
make the tool fairly usable...)

I also think Together is superior to Rational. It is probably the best
tool on the market despite my criticism.

Well, now... I just have to see if it is worth the money... Not only is
Together expensive, but on top of that, you've got an expensive machine
to buy. It gets *really* expensive... Worse thing is not the money...
buying a new machine is a hassle, you need to reinstall everything...
and then, you have to consider the fact that Together won't run on your
older machine... that's not very interesting... The question then is
whether a cost/benefit analysis would lead you to buy such a product... 

Right now, Linux is in a bad spot. The only good developer's tool beside
kdevelop are written in Java and most of them are mostly good at
programming in... Java. JBuilder for example is really a great tool.

So, this leads me to look forward at how kdevelop will evolve...

I see myself in two years using kdevelop quite a bit more than I do now
if kdevelop really meet my expectations (right now, I'm really only
using makefiles with a good editor {jEdit} because most of my work isn't
GUI bound). By that time, Together might be a good companion.

> 
> On Mit, 29 Nov 2000 Daniel wrote:
> > I would not recommend Together 4.1 which I tested last week.
> >
> > Since it provides both way engineering for C++ and Java, it seemed
> > promising... And I use both languages extensively, so it is really
> > interesting.
> >
> >
> > I've known about Together since version 1.0 I think. Back then, I was
> > trying to run it under JDK1.0 with a Sun Solaris box... it was slow...
> > so slow...
> >
> > Now, it runs fast. I would say it isn't there yet as far as speed is
> > concerned, but given all the work it does, it is fine. Those of you who
> > have a 1,000 Mhz computer will think it is ok.
> >
> > The problem is elsewhere. It is a dangerous tool. The editor doesn't
> > backup your changes and if an extension is thrown (anything, but often
> > and OutOfMemoryException can occur on modest projects...), then, it will
> > rewrite your code... leaving out the pieces it could not get at due to
> > the exception.
> >
> > Finally, together, on a modest project (say 20-50 classes) will easily
> > require 200 Megs on its own to run smoothly. I really mean it. That's
> > how much memory you won't have for other applications. The thing is a
> > monster... it loads stuff up and never releases the memory...
> >
> > Obviously, the tool has been optimized for speed, not memory usage!
> >
> > So, if you have a 1,000 Mhz computer with 512 Megs of ram and you are
> > backing your files manually ever 15 minutes... it is fine.
> >
> >
> Hi Daniel,
> 
> thank you for letting me participate in your experiences.
> I did some testing as well and want to let you share.
> (Please let me know if I'm to much OT - I know it is a kdevelop list!)
> 
> 1) Rational Rose:
> I testet Rational Rose (currently only on WinXX but the Linux Beta is coming
> out this year).
> What I don't like is the way they modify my C++ code. It is covered with
> comment lines containing ID-numbers etc. . It gets hard to read. So I think if
> you decide for Rational Rose you have to swich your complete development. On
> the other hand it is fast and very stable.
> 
> 2) Together 4.2
> It's based on JAVA - means running on Linux. Positive!
> It doeasn't modify my code, it's like kdevelop parsing the code to get the
> classes when reengineering. No ID-comments etc. .
> The style of headers, code and comments is perfectly fitting with mine (and
> kdevelops) when adding methods etc. . The docstyle is JavaDoc :-)).
> The initial diagrams are much better than those of Rational - better
> arrangements. Sequence diagrams are excellent, you can even edit them and
> update the code from your graphical modifications!
> But now the odds: Memory consumption is horrible and after a while it eats up
> all your resources. I think it has a lot of memory leaks.
> Because of its nature (Java and continuous class parsing) it is slow.
> OK, I tested it on a 800 MHz/128 MB. I'll see how it runs on my new
> 2 x 800 MHz/256 MB which is (hopefully) coming tomorrow :) .
> I recognized that it crashed my X-server several times when run on another
> machine and the display exported to my machine via X11. This didn't happen with
> local display.
> 
> Conclusion: Both tools are comparable in price (and much to expensive!!!)
> I dont go with Rational Rose because I don't want to do graphical software
> engineering but want to stay closer with the code.
> Together seems to be OK for me since I don't want to use it continuously to
> create code. I want to use it only for making diagrams and keeping the overview
> in my multi developer projects.
> 
> I hope this all is of some use for the developer comunity on the list here.
> 
> - Werner -
> 
> --
> --------------------------------------------------------------
>                     ALC Computertechnik GmbH
> 
>         \_       \_          \_\_\_     Kaiserstr. 100
>        \_\_      \_        \_           D-52134 Herzogenrath
>       \_  \_     \_        \_           Germany
>      \_\_\_\_    \_        \_           Fon +49 2407 95810
>     \_      \_   \_        \_           Fax +49 2407 958139
>    \_        \_  \_\_\_\_    \_\_\_     http://www.alc.de
>    ===========================================================
>    Werner Modenbach                     modenbach at alc.de
> --------------------------------------------------------------
> 
> -
> to unsubscribe from this list send an email to kdevelop-request at kdevelop.org with the following body:
> unsubscribe »your-email-address«

-- 
Daniel Lemire, Ph.D.

http://www.ondelette.com/

-
to unsubscribe from this list send an email to kdevelop-request at kdevelop.org with the following body:
unsubscribe »your-email-address«



More information about the KDevelop mailing list