> >
> >> I know it's a huge change but here are my 2 cents:
> >> - Boost is slow

Would you be able to point to somewhere that provides some comparisons?  I
was under the impression that since templates are basically a compile-time
thing, the compiler can optimize a bunch of stuff at the time of compile,
and maintain compile-time type safety.  Qt does a bunch of run-time
type-safetying, which, by definition, increases runtime.  This is the
perennial problem of run-time optimization or program space optimization.

> > That is true but I think it is mostly Boost.Signals that contributes
> > in this way to the size of Wt (since for every combination of types,
> > new code is instantiated). But I do not think that warrants going
> back
> > to a moc approach, since this is the single most criticized design
> > feature of Qt, which was very defendible at the time but no longer so
> > given current state of C++ (and compilers).
> 
> To moc or not to moc is not really an issue, IMHO.

I would put my two cents and say that I dislike moc'ing.  I don't use those
parts of Qt that require moc'ing...  Which does leave out a bunch of fancy
stuff.  When boost doesn't have something, I've used some stuff from QT.
And if I can't use stuff from there, I'll move onto POCO, et. al.

> 
> >> - Qt provides lots of features Boost does not even dream of having

I suppose QT is the open source world's answer to Microsoft's CLI (aka
.Net).  


> >
> > Now that is true -- Qt is without doubt the single library that would
> > be the most valuable dependency for having in Wt. Although it cannot
> > replace Boost.Signals in my opinion, it could possibly replace
> > anything else that we use of boost, and provide many new features:
> 
> Why can't Qt signals and slots replace Boost.Signals? IMHO Qt signals
> are better than Boost signals: they can be queued, introspected, can
> ignore parameters, etc (see
> http://web.archive.org/web/20070703100120/http://scottcollins.net/artic
> les/a-deeper-look-at-signals-and-slots.html
> )

Even after reading through that, I am still of the belief that compile time
type checking and optimization has a valid position in program composition.
And it high utilization web sites, at the cost of extra program size through
template instantiation, good run-time does have its merits.

> 
> >  - raster image paint devices for WPainter,
> >  - QSvgRenderer for rendering an SVG file on a WPaintDevice (and
> > implicitly converting SVG to VML or canvas)
> >  - more unicode transcoding support besides UTF8,
> >  - database access layer,
> >  - application-level networking protocols (email, ...)
> >  - scripting
> >
> > It remains the question whether Wt should make the decision to adopt
> > Qt, or whether we should wrap selected Qt features as an add-on
> > library that integrates well with Wt (e.g. for some of the features
> > mentioned) or whether it presents an opportunity for Wt application
> > developers to leverage Qt features where needed.

Adopting it with wrapping, and if at all possible, without moc'ing, would be
of value to me.

> 
> > And, allthough a LGPL license allows closed source applications, it
> > does prohibit static linking, which still results in a much larger
> > flash footprint in embedded systems (since static linking allows the
> > linker to dismiss unused symbols), and this should be kept in mind as
> > well.
> 
> No need for dynamic linking. It is generally agreed the LGPL allows
> static linking as long as you provide the compiled source (i. e.
> object files) you would need if you were to re-link your application
> to a newer version of the LGPL library (Qt, in this case).

Whatever the outcome, static linking is preferred for me.




-- 
Scanned for viruses and dangerous content at 
http://www.oneunified.net and is believed to be clean.


------------------------------------------------------------------------------
This SF.net email is sponsored by:
SourcForge Community
SourceForge wants to tell your story.
http://p.sf.net/sfu/sf-spreadtheword
_______________________________________________
witty-interest mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/witty-interest

Reply via email to