Re: "What's this"

[Sasha_Vasko osca state mo us]

> On Mon, 7 May 2001 Sasha_Vasko osca state mo us wrote:
>
> > Clients should not be bothered with displaying help, unless you want to
> > bloat all of them with basically identical help rendering code.
>
> Clients are the only pieces that know how to best display the help. If
> someone wants to implement the paperclip (whether you like it or not,
that
> case needs to be possible), doing it in the client is the only sane way.

While using proposed selection mechanism, it would be possible for client
to
respond with some special answer, indicating its desire to "Take over" help
displaying. I would estimate, thou, that 95% of Unix/Linux/open source
apps will not want to do that. It really is kinda stupid to write a man
page,
and then code huge piece of code into your application to display basically
same information. I understand that this is different for BIG applications
like Web Browsers/Word Processors, but we should not be thinking solely
about those, and if there is a simple way of implementing generic
algorithm,
that automagically supports ALL unix applications, without changes required
to
most of those - we should go with it. Using huge library of man pages is a
good thing (TM).

>
> If client wishes to use a standardized-help-display-mechanism to show the
> popup help, then it is entirely possible, but that's a separate problem
to
> be solved - let's not complicate the problem further than it has to be.
> Right now the only thing to do is find out whether a client has help
> available for a widget (to provide cursor feedback) and ask the client to
> provide that help...

Selections mechanism is very simple to implement, and most of those BIG
apps
that needs it, should already have some selections management code, since
things
like clipboard are implemented this way.

On the other hand standardizing some inferior ad-hoc solution is a nasty,
way to go since it will provide a very limited solution (you don't have
2-way
communications, so there is no way to find out if app has acknowledged
request, and therefore you can't implement any fallback techniques ). And
we'll need to live with it till the rest of our lives, and ppl will be
pointing fingers, blaming us in very harsh words, and we'll be tearing our
hair
out, and there will not be a way back, etc. , etc., etc.
Sooner or later some 2-way communications will have to be implemented, so
why not do it right the first time. It's not like if we don't come up with
something today - world collapses.

>
> So, I'm in rough agreement with Havoc on this one,
> -- Elliot


Regards
Sasha Vasko