Text 58, 143 rader
Skriven 2006-04-11 16:34:00 av Michiel van der Vlist (2:280/5555)
     Kommentar till text 37 av Robert Couture (1:229/2000)
Ärende: FD & -Unpublished-
==========================
Hello Robert.

10 Apr 06 20:50, you wrote to me:

 MV|>> He never used Irex himself. All he has is what *others* told him.
 MV|>> For him it is hearsay.

 RC> I suppose that is true.

Which was my point.

 RC>>> No, sorry ... it is very *broken*  If it does not run at *all* it
 RC>>> is broken.

 MV|>> But it *does* run. It has been running here for three years.

 RC> It does not run under all the platforms it is released under.

I can conform... eh confirm ;-) nor deny that it does not work under Linux.
never tried it.

Maybe Charles' mistake was to try to cover so many platforms in the first
place....

 RC> There is no excuse for no release of a version for Linux compiled to
 RC> work under newer kernels.

I can think of at least three valid excuses....

 RC> From my understanding, all that is requried is recompiling using the
 RC> libraries from a newer kernel release.

Your understamding may be wrong...

You know I have learned to be not so quick with statements like "all it takes
is a simple this or that", especially if it is something I can not do myself.
My experierence is that in many cases what looks like a simple thing turns out
to be a major operation if possible at all.

 RC>>> Hopefully Charles can be convinced to open source it should he
 RC>>> decide to stop working on it himself.  That would be best.

 MV|>> Then it still would take someone willing and competent to pick it
 MV|>> up.

 RC> If it is open source then it only takes a bit of effort on a few
 RC> people.

Maybe. Or maybe not. Maybe the source is such a mess of spaghetti programming
that no one can make heads or tails of it. Maybe you will dearly regret ever
having asked for it to be released as within a few days a hacker will have
discovered a dozen vulnerabilities and written exploits for them. Opening
closed source is no a panacee your know. Publishing the source may very well be
the mercy shot...

I am not suggesting Irex falls in this class, but I also suspect the reluctance
of many a programmer to release the source of his product is that he knows it
isn't up to standards and does not want to expose himself as a bungler.

 RC>>> If not that, at least pass it / sell it to someone who wants to
 RC>>> develop it more.

 MV|>> After what happened to InterMail, I am not looking forward tp
 MV|>> have the same happen to Irex. Peter Stewart sold it to Dale
 MV|>> Barnes, who promised all sort of things, but eventually t he too
 MV|>> left FidoNet taking the source with him.

 RC> Well if I could afford it, and Charles would sell (as a last option --
 RC> I would prefer it was given away <g>) I know I would release it as
 RC> open source.  At minimum I would make it available on CVS to those
 RC> interested in tweaking it and applying bug fixes, et al.

 RC>>> If it comes down to sell, maybe the Fido community could pony up
 RC>>> a donation for a fund to purchase it for everyone :)  Or
 RC>>> *something*

 MV|>> Hmmm... I already paid once for Irex. why should I pay again when
 MV|>> there are freeware alternatives?

 RC> I understand ... I was thinking of the betterment of others with the
 RC> help of a some.  Irex is a good program. I would rather not see it
 RC> die.

Is it? A good programme? Maybe the reason you can't get it to work under Linux
is because when we all get to look at the source it truns out that it isn't as
good as you think it is. Maybe the reason it did not work to your satisfaction
under Win32 is that it *is* fundamentally broken as you claim. Maybe it is
broken beyond repair and the best thing to do would be to put it out of its
misery.

I am not saying that it is, just that it could be...


 RC>>> I just want to use it again :)

Actually I find your attitude a bit puzzling. You claim that it is broken in
various ways and yet you want to keep it alive and use it again....

 MV|>> I am not married to Irex. So far it suits me. If the day comes
 MV|>> that it does not suit me any more, I will change to something
 MV|>> else. Something free....

 RC> The only thing that comes close to Irex' features is Taurus.  But it
 RC> is Win32 only.  :(

So why not give MBSE a try. The author is an active member of FidoNet and it is
Linux...

http://62-221-196-62.dsl.fiberworld.nl/


Now regarding this "just a recompile" thing: I have my doubts about trying to
cover all platforms. yes, ever so often one can get something to get to work on
another platform by a recompilation. But... to get it to work *real* good you
have to do a lot more than that. you follow the Golded area. Why does the Win32
version of Golded performs poorly under Win9x? I think basically because it
just a recompile of the DOS version.

To really get an application to perform well, it has to be optimised for the OS
and for that the programmer needs to know the OS fairly well. No one can be an
Expert in all OS's however and that is the problem with these multi-platform
things.

For Golded Win32, one can say that the command line interface on Win9x is the
culprit. I say if that is so, then a command line utility is the wrong way to
go for a Windows application. A *proper* user application for Windows should
use the GUI rather than the CLI. That would also make it possible to *properly*
implement utf-8. That is without having to have to deal with this codepage shit
etc....

But if it was completelY rewritten that way, porting it back to another OS
would be almost impossible. This multi-platform approach does not work very
well...

Alex Shakhaylo has written Taurus for Windows. he is not contemplAting versions
for another OS AFAIK. Michiel Broek has written MBSE for Linux. He is not
comtemplating a Windows version AFAIK. I think they both are doing it right.


Michiel

--- GoldED+/W32-MSVC 1.1.5-b20060315
 * Origin: http://www.vlist.org (2:280/5555)