Re: comment about gnome architecture

[Alexander Larsson <alexl redhat com>]

On Tue, 2003-12-16 at 21:01, Greg Breland wrote:

> If you were referring to writing the implementation of a spec that
> requires software, fontconfig for example, then yes, a bad spec could
> mean that the software fails.  But how is this bad?  Do you think the
> software would have succeeded if there was no spec?

Yes, if there was no spec then the software could have been made good
intstead of being forced to be bad.

In fact, the example you bring up is an excellent example of a really
bad specification, the XLFD (X Logical Font Description). This
specification was made in order to make some sense of X11 fontnames,
which are otherwise just opaque strings. If fontconfig was to use this
specification, it surely would have failed. Instead the developers chose
to ignore the specification and implemented their own font naming
system. However, we had the bad old spec for many many years before
anyone tried to replace it, and we will still have it for many more
years...

Now, by virtue of being the only serious contender to a real font system
for X11 the fontconfig naming system could be considered a
"specification" now, but thats really just stretching the idea of
specifications so that any implementation is a specification, making the
whole discussion useless. Fontconfig is not a specification, it is good
software.

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 Alexander Larsson                                            Red Hat, Inc 
                   alexl redhat com    alla lysator liu se 
He's an old-fashioned flyboy stage actor in drag. She's a provocative punk 
hooker with someone else's memories. They fight crime!