Re: ICNS as default format

[Gregory Casamento]

Riccardo,

When PNG support was first added, there were endian issues there, so
I'm not certain how implementation or support of ICNS differs in the
amount of support we should give it especially in light of the fact
that it's widely used on Mac OS X.   If there are bugs with handling
endianness, then we should fix them if they are on our side and tell
the author of the library, if they are on his side, if you happen to
be using the library..  I actually don't think the libicns library has
this problem at all.

When using the library, the only feature missing, as I recall, is the
ability to select the appropriate image based on the image size
needed.  This may have changed recently.

Additionally, if you're talking about the external dependency...
regarding it's portability.  Portability is a function of how many
platforms to which the library has been ported.  I've seen packages in
the following distros for it, this is not a complete list, I'm sure:

* Debian 6.0
* Fedora
* SuSE
* CentOS

I'm sure there are others.

Also, without the library, using our drop in replacement, I'm not sure
how you can say that the format isn't portable when Fred's
implementation compiles on every single platform I've ever tried it
on.

Where ICNS currently fails is the ability of image editing programs to
create them which, I suppose, is a major drawback to the format.

Later, GC

On Thu, Jun 9, 2011 at 5:14 AM, Riccardo Mottola
<address@hidden> wrote:
> Hi,
>
> I agree with Fred. Supporting icns is fine for compatibility reasons, but it
> proved a format to be rather difficult to implement and a bit unportable.
> The current implementations has endianness and portability probelms (and the
> other one based on an external dependency was only marginally better).
>
> Tiff is also a very well supported format: I love to be able to create icons
> and imagery in GIMP, photoshop or whatever else. I still find MacosX
> uncomfortable in that direction, especially since they removed even backward
> compatibility.
>
> Riccardo
>
> On 06/09/2011 10:46 AM, Fred Kiefer wrote:
>>
>> On 09.06.2011 01:22, Gregory Casamento wrote:
>>>
>>> Just a quick thought....
>>>
>>> I'm wondering if we shouldn't consider using icns as the default for
>>> icons.  Once the icns implementation is completed it should, in
>>> theory, adjust to the optimal resolution for the given size.
>>
>> Not sure what the benefit would be. The tiff format can also hold multiple
>> images at once and the tiff library we use is more common than the icns one.
>> And our own icns loading code is rather incomplete. I should know that
>> having written most of it.
>>
>> Just to quote from that file:
>> /*
>>  The following code is a drop in replacement for libicns. It may be used
>>  when the library is not available or unsuited due to its licence
>> (Currently
>>  GPL 2). This code was mostly build based on the documentation found at
>>  http://icns.sourceforge.net/apidocs.html. It also includes icns decoding
>>  ideas based on code in mySTEP.
>>  Only limited formats are implemented and some errors still exist.
>>  */
>>
>>
>> Fred
>>
>> _______________________________________________
>> Gnustep-dev mailing list
>> address@hidden
>> https://lists.gnu.org/mailman/listinfo/gnustep-dev
>
>
> _______________________________________________
> Gnustep-dev mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/gnustep-dev
>



-- 
Gregory Casamento - GNUstep Lead/Principal Consultant, OLC, Inc.
yahoo/skype: greg_casamento, aol: gjcasa
(240)274-9630 (Cell)