Re: [gobo-eiffel-develop] gobo ecf

[Colin P. A. <co...@co...>]

>>>>> "Martin" == Martin Seiler <se...@et...> writes:

    Martin> Dear Eric,

    Martin> I just want to express that I feel very unhappy about how
    Martin> you acted in the ECF discussion over the past year.

I on the contrary fully support Eric's stance. 

    Martin> I personaly hated Ace since I had to deal with it in my
    Martin> first year at ETH.  I _like_ ECF. I do not force you to
    Martin> like it...

I don't like ACE either - but that's because I am quite unfamiliar with
it, having always used Gobo's XACE format - which is simple, and
portable across all known Eiffel compilers.

    Martin> I respect your opinion.  But you do not respect
    Martin> mine.

How can you make such an assertion?

    Martin> I want ECF. I want ECF support in _all_ gobo tools.
    Martin> This is the need of your users. And I think it is time for
    Martin> pragmatism.

It is not the need of Gobo users. Gobo is for portable Eiffel
tools. ECF is not a portable tool.
It has previously been suggested that the syntax of a configuration
file does not matter - using the configuration library is all that is
needed, and it could be made to support ACE. But no such support has
been offered.
Eric has put a lot of work into writing support for parsing ACE files,
and also for the XACE tool, which provides a portable configuration
file format, and is easily extended.

    Martin> Not having support for it caused me and others already
    Martin> enough pain.

Which is quite predictable, and entirely the fault of Eiffel
software. We already had a portable XML configuration file format (I
entirely agree with Franck Arnaud that XML is highly suitable as a
syntax for such tools), which could easily have been extended to meet
Eiffel Software's needs, but they chose to ignore this possibility,
and developed their own format, without public consultation.
You are shooting at the wrong target.

    Martin> And now you tell me that actually code
    Martin> exists but you don't release it?  This is a punch in the
    Martin> face of your users...

Hardly. The code concerned is copyright by Eric's employer, and is
not eligible to be included in any open source project, unless the
employer should decide to do so.

But I have used this code (I work for the same firm as Eric). While it
allows parsing an ECF file, it doesn't overcome the problems with ECF
- I know - I have suffered from them.

    Martin> You act very much
    Martin> against the way open source projects work.  I simply feel
    Martin> that gobo is _not_ open enough.  There was more than one
    Martin> attempt of people to add this support to gobo, but you
    Martin> blocked it.  And now you did it again.

This is not true. No-one has offered to donate any software that would
generate ace or xace files from an ecf file, which is the only "pragmatic"
response to the mess that Eiffel Software have got us into. 
But even that action (which Eiffel Software has a moral responsibility
to do as an interim measure) won't overcome the problems with ECF.
-- 
Colin Adams
Preston Lancashire