[openbeos] Re: ShowImage Bloat

  • From: Adam K Kirchhoff <adamk@xxxxxxxxxxxx>
  • To: openbeos@xxxxxxxxxxxxx
  • Date: Mon, 19 Jan 2004 10:18:46 -0500 (EST)

On Mon, 19 Jan 2004, Simon Taylor wrote:

> > I've been following this entire ShowImage discussion with a little
> > bit of
> > interest.  Just a few quick points:
> >
> > 1) Disk space may be cheap, but that's no reason to waste it.  Anyone
> > who
> > has worked with large media files knows that every byte counts.
>
> Hmm, OK. I don't think thats a very good argument when we're talking
> the order of a few kB. The user is more likely to say - "ow, I've run
> out of space - better delete that app I don't use any more" than to say
> "noooo I've run out of space - and it's all the fault of whoever added
> the slideshow to ShowImage!"

If we say it's OK for ShowImage, what do we say when it starts happening
with every other app and/or utility.  Suddenly that few kB starts adding
up fast.

> When I'm talking about a dislike of bloat, it has much more to do with
> the experience of using the app (more features do not necessarily
> enhance that) than the exact size of the binary.

I think that both aspects are important enough to take into consideration.

> > 2) The fact that this debate has gone on for so long indicates that
> > there
> > obviously needs to be a clearer definition of what OpenBeOS R1 is
> > aiming
> > for...  Is it to be an opensource implementation of BeOS R5, an
> > opensource
> > implementation of Dan0, or are we aiming for something more?
>
> My personal view is that R1 is a reimplementation of R5. That does not
> mean that it is a clone of R5 - but a reimplementation of R5's API and
> general features. That means moving networking to the kernel is a good
> move (same basic API for coders, better implementation), along with the
> host of other improvements being made for R1. That definition also
> rules out things like multi-user support for R1.
>
> Something like adding features to ShowImage is certainly within the
> realm of R1. For me that would mean supporting more file formats
> (multipage tiffs is one example), and zooming support.

If we're talking about just reimplementing the R5 APIs, I agree.  If we're
talking about reimplementing R5 (which is different than reimplementing
the APIs), then I disagree.

> [different mail]
> > - Michael
> > Print Kit leader who thought improving ShowImage could not hurt...
>
> Michael (and Michael :-)) I really do appreciate the work you have
> done. This thread has rather glossed over some of the very useful
> additions to ShowImage that you have made. Thanks.

<snip>

I agree 100% with everything else you said :-)

Adam


Other related posts: