[Glade-devel] libglade usage

[garnacho www tuxerver net (Carlos Garnacho)]

Hi Ishan,

On Mon, 12 Jan 2004 21:05:59 +0530, Ishan Chattopadhyaya wrote
> Dear Dragon and Murray,
> > > > I've seen entirely too many suggestions on this list along
> > > the lines
> > > > of "I don't use that feature so delete it."
> > [snip]
> > > Just a thought... did you mean removing *source code
> > > generation* in Glade-3 by this?
> >
> > A truly useless and bad feature should be removed. If you disagree with
> the
> > glade maintainers' ideas of what is bad and useless, then nothing stops
> you
> > from maintaining glade code-generation as a separate module. If you will
> not
> > do that then you don't care enough. If you can not do that, then you might
> > not be qualified to judge the issue.
>
> I'm not so sure that source code generation as in glade-2 is 
> 'useless' and a 'bad feature' as u put it. The point that I feel is 
> that if it's not upto the mark or not well maintained, efforts 
> should be made by somebody ( who is willing to spend time with it )
>  to re-write the feature (as you mentioned).

I think that if code generation were the right solution, the whole GNOME 
project would use it, but here's a list of disadvantages that mostly anyone 
could quickly find:

- Could mix code with UI to the point of making things like HIG impossible to 
be applied (every UI change could be a pain)
- the code generation would be bound to a single language, making harder to 
people to colaborate in their favourite language. OTOH, there could be 
language generators for every language that GTK+ supports, but that's IMHO 
widely out of the objectives of glade
- Let's assume it, code generation never is perfect, it always generates code 
that you don't mind/you'll don't use/you don't even care and will generate it 
in some way that may be more hard to understand or even less pretty, that 
decreases the code quality and sometimes would make it unmaitainable

> Just a suggestion, why not output/generate code that is compatible with
> LibGlade by default with glade2. And for the users who want better coding
> styles/approaches can alwayz use ext. XML parsers. To put it in 
> another way, why not somebody makes a *good* source code generator 
> and that generator be bundled with forthcoming glade3 releases by default?

making a *good* code generator is not *easy*, and if people says about 
deprecating it, I think that this is because they see many more disadvantages 
than advantages for still working on this

> When the GUI of glade can be be used for source generation, it's 
> more user friendly. Of course, people can always use the terminal 
> for that if they need. I think although it's the job of the IDEs to 
> generate code, why not keep the option to the users to use it as 
> standalone as in case of source generation?

I don't think in glade as an IDE, and anyways I'm pretty eskeptical to the 
code generation shipped in some IDEs

> The reason i feel a source generation option should be there is that 
> many times I (and maybe other users) might feel the need to 
> _quickly_ check the functionality of the interface in runtime. 

I really think that it's more quickly to save the glade file and to run the 
program that uses it through libglade, so no recompilation is needed to test 
the UI, can be done with your favourite language, etcetera...

> (after all glade is pretty much a rad tool). So, I don't feel it's 
> harmful for the users in any way to let them use what they want. But 
> since source code generation is a deprecated option, proper caution 
> (such as an alert box) can always be given to the user before he 
> uses such a feature (that is if such a feature is indeed 
> implemented/allowed to remain unaltered).
>
> BTW, I don;t want to influence anyone's decision by saying all this, 
> but just letting the people know my thoughts on this.

me too :)

> > nothing stops you
> > from maintaining glade code-generation as a separate module. If you will
> not
> > do that then you don't care enough. If you can not do that, then you might
> > not be qualified to judge the issue.
> That's a *rubbish* assumption. There might be many users (glade-2) 
> out there who still use source code generation. Now if they want 
> that feature but don't want to maintain a new module for that, does 
> that mean that they dont care about such a feature? Not everybody 
> who cares can devote time to do that. And does it mean that people 
> who 'cannot' maintain such a project are not qualified to judge? 
> Acc. to me, the users are wise enough to judge what they need. But 
> even if they are not, they are free to suggest the people in the 
> decision making about what they feel.
>
> Regds,
> Ishan
>
> _______________________________________________
> Glade-devel maillist  -  Glade-devel lists ximian com
> http://lists.ximian.com/mailman/listinfo/glade-devel


--
Carlos Garnacho <garnacho tuxerver net>
                <carlosg gnome org>