[news.eclipse.tools] Re: Why should I use SWT?

[mcq@xxxxxxxxxxxxx (Mike Wilson)]

I am part of the SWT team. :-)  I understand Mike Walker's concerns
about this, and I *definately* do not want to start a holy war on
either side of the issue. I can say that we are quite confident in the
decision to go with SWT over Swing.

David Whiteman makes several good points...

Widget performance, especially across a wide range of targets, is
definately important to us. The kinds of tricks Swing uses to get
acceptable performance on Win32 just aren't available on several of
the places we want to run.

The issue of emulating the look and feel of a platform versus just
"being" the look and feel of the platform is also very important. If
you count the time to being *done* a port to a particular platform, as
the time it takes to get to a point where you no longer 'seem clunky'
to the user rather than just the time to get *something* going, then
SWT is definately easier to port than Swing. ;-) You also can't
respond to changes in the platform's look and feel as a result of
operating system changes unless you are built out of the native
widgets. And finally, by being that close to the platform you get
access to all of the support that the platform provides for things
like BIDI, accessability, etc. No matter how many person years Sun has
put into this, you can safely assume Microsoft will put more. This
will be true for every platform where the vendor wants to be
successful in the modern world. 

There's one other important point which David missed: SWT is something
which can be made part of the open source offering, with all of the
benefits that that entails.  I'm not worried about porting SWT to
multiple platforms;  I have already got several other instantiations
in house (*sigh*. I'd like to tell you which ones, really I would!)
and by enabling the community at large to make new ones, I'm sure
there will be many more. In addition, if SWT needs to evolve to better
support the things being done in eclipse (and it's other consumers),
then we (the community) can make it happen without being constrained
by whether or not it fits with Sun's vision.

Just so you know, I'm unlikely to get involved further in this
discussion. I don't feel it's my place to be the "shining defender of
SWT". I like it. It works. 'nuff said.

McQ.

On Tue, 10 Jul 2001 09:04:40 -0400, "David Whiteman"
<david_whiteman@xxxxxxx> wrote:

>I'm not part of the SWT or Eclipse team, but I'll tell you a few things that
>I find compelling about the SWT story (this is my personal opinion and not
>any official stance by OTI or IBM):
>
>* Native widgets will always perform better than emulated ones.  This is
>most important in very small embedded environments.  Sure, processors will
>get faster and memory is cheap, but we are constantly trying to fit
>computing power into smaller and smaller boxes.
>* While there are some impressive efforts to try and emulate the native
>look-and-feel, it never seems to be done quite right.  Users can tell the
>difference, and your interface will seem clunky.  In addition, when a new
>LAF comes out for your favorite OS and you are using emulated widgets, you
>then have to wait for that LAF to be developed and the kinks worked out
>before you can actually use it.
>* When you use native widgets, you can leverage tools for automated testing
>(such as WinRunner) and leverage peripherals (such as the wheel mouse) with
>no added support necessary
>
>My .02,
>David
>--
>------------------------------------------
>David Whiteman
>david_whiteman@xxxxxxx
>------------------------------------------
>"Mike Walker" <michaelw@xxxxxxxxxx> wrote in message
>news:9ickh6$h4s$1@xxxxxxxxxxxxxxxx
>> We are just about to embark upon a new Java project based on the Eclipse
>> IDE. We will be building Eclipse plug-ins to add our functionality to the
>> IDE. I am, however, deeply concerned that the Eclipse project has made the
>> wrong decision to go off and invent a new set of widgets - the SWT -
>> instead of using Swing.
>>
>> I have read through the document explaining how the SWT works, and I can
>> see a lot of thought has gone into making a good design, but I still don't
>> see the justification for the SWT as opposed to using Swing.
>>
>> My concerns are as follows:
>>
>> 1) Platform neutrality - do you claim that the SWT will be as platform
>> neutral as Swing? I can see how you can make the basic functionality work
>> the same on all platforms but can you really make *all* of the behaviours
>> of all the widgets work the same over all platforms? I can speak from
>> personal experience - I worked on porting the AWT to OS/2 for a couple of
>> years and the amount of pain we went through to get every little thing to
>> work just right was incredible - e.g. triggering events at the right time,
>> input focus handling, sizing correctly, font handling, code pages, etc.
>> OS/2 is very similar to Windows and yet we had a terrible time with these
>> issues and others. I can't see how the SWT will be able to avoid these
>> very same problems.
>>
>> 2) Functionality. We have requirements (in IBM) to support multiple
>> languages and accessibility functions. Does the SWT provide that support?
>> Do you have Unicode support? Do you support BIDI langauges (Hebrew and
>> Arabic?), or the Asian input methods? If not, how long before that support
>> is there? All this is supported in Swing (after many person-years of
>> effort) today.
>>
>> I guess that my main problem is that I can't see the justification for the
>> SWT. I came only surmise that you are concerned about performance since
>> that can be the only advantage. However, with the speed of today's
>> processors, cheap memory, and advances in JIT complier technology, this is
>> becoming a much less of an issue.
>>
>> Am I missing something? Can someone spell out to me the reasons why the
>> SWT was invented?
>>
>> (Note: - I really like the idea of Eclipse and am looking forward to using
>> it - it is just this one area I am very concerned about).
>>
>> Mike
>>
>
>