[cairo] Re: Chaco DisplayPDF with Antigrain Internals

[Ross McFarland]

On Fri, 2004-08-20 at 17:16, M. Evans wrote:
> Ross McFarland wrote,
> 
> > they're normally not nice for libraries as they make language bindings
> > very difficult.
> 
> And you expect people to infer this statement from three bang marks...wow.

i don't expect that people people infer it, i was just giving what
reasoning i could come up with (guess) for it.

> This rationale for C is common, and false.  Of course, no one expects or 
> wants Cairo to change from C.  (Although a safe-C treatment like CIL 
> makes sense.)
> 
> A price is paid for the supposed benefits - slower development, bugs, 
> low-level API mentality, etc.

i won't argue with that really. there are quite a few problems with
cairo that stem from C, though there hasn't been any major effort to
address these types of issues yet (from what i can tell.) 

>   The ease-of-binding claim is dubious. 
> For example, an API which exposes C pointers all but guarantees that it 
> will only be linked to C/C++.  Each language has to deal with the 
> pointers in its own unique way, through C glue code.  I don't call that 
> easy! 

i guess i mean more easily bound than other languages. C/C++ can be
bound to perl without too much headache, but i'm not even sure where i'd
start with binding something written other languages languages. likely
the bindings would still have to be done with C/C++ and that binding
would have to do whatever was necessary to get from whatever language it
was developed under to C and then to the bound language.

>  Now, a TCP socket which accepts string commands, that is easy 
> from most any language - not that I recommend it, just for comparison.

now you're talking about .net/mono/bonobo type things. while they could
be done performance (as i assume you're already thinking) would be
questionable.

> Overall what Cairo needs is a feature matrix chart showing what exists, 
> what exists partially, what doesn't exist, what will never exist, what 
> might exist, and in what time frames.  I sometimes feel like I'm talking 
> to Microsoft, for all the future-ware that is being promised.

sure that would be a good idea. i wonder what the status of cairo really
is sometime, and eve more what the projected dates are for when it will
be full function and such.

> Cairo is a great project, but not the only one, so it would be nice to 
> get an overview "at a glance" that disambiguates future-ware from 
> present-ware from nearly-there-ware from never-ware.
> 
> Another thing.  What makes Kiva / Chaco appealing (to me) is not only 
> the BSD license, but the fact that it's being used in real applications, 
> today.  It would be nice to see some of those apps on the web, too, when 
> they come (as opposed to unit-tests and mock-ups).

from what i've seen cairo seems to be pretty early in its lifetime. i
don't think anyone is claiming that it's full function and ready for
prime time use. (i would like to see someone's thoughts on when it might
be or when they want it to be.) 

-- 
-rm
http://www.neces.com/