Re: [PuTTY] : status

[Mohammed Elzubeir <elzubeir at arabeyes dot org>]

On Tue, Jun 18, 2002 at 06:44:40PM +0100, Simon Tatham wrote:
> Mohammed Elzubeir <elzubeir at arabeyes dot org> wrote:
> 
> Have you told them that?

We've had endless discussions about this on the XFree86-i18n list.
http://www.xfree86.org/pipermail/i18n/2002-February/thread.html

Search for bidi, and you will find more than half of it is on this
subject ;)

> What worries me about this situation is that there are these two
> schools of thought and neither of them shows any public recognition
> of the other. An independent terminal emulator developer such as me

We are not in a position to recognize or not recognize such a thing.
It is more of it being the other way around ;) Our efforts have mainly
been to bring Arabic support to what we see as vital applications. We
have had several issues with XFree86 in particular (1. font related,
2. xterm related). Unfortunately, until now we have failed to come
to a satisfactory conclusion on either subject.

> has no sensible way to work out what the consensus opinion is, or
> who's right, because I haven't seen any counterarguments that you've
> published about their position or that they have about yours.

Someone from XFree86 did recommend that we do just that. I think it is
high time that someone does in fact write such a paper. 

> Making the terminal emulator's behaviour simple means that the
> terminal emulator itself is simpler, _and_ that every application
> which needs to predict the behaviour of the terminal emulator is
> also simpler.
> 
> It's true that for some applications simplicity in the terminal
> emulator means more complexity in the application, but for others
> it's the other way round.

That is true, and for that same reason Arabeyes went on to patch the
more demanding applications (e.g. vim). Of course our numbers are
limited and so are our resources -- so we can only do so much at 
a time ;) And with so many contradictory opinions on how things should
be implemented, it doesn't make it any easier.

> Er, adding optional modes of behaviour _doubles_ the
> unpredictability, because your application then has to guess what
> mode it finds the terminal in. It doubles the testing work, as well.

Noted.

> Suppose you're writing an editor. You display some text. Now the
> user moves the cursor to a particular position, and types some more
> text. Where in the document does that text end up?
> 

mlterm+vim seem to work pretty well at that. VIM currently does not
support bidi but does support Arabic shaping, and mlterm does all
the bidi work. The cursor does tend to jump around when you move it
over LTR then RTL text, but it is something you get used to and then
learn how to work with it.

> Yes, you're mistaken. I said _by default_. Whatever xterm does,
> PuTTY will also do in its default configuration, because otherwise
> it will confuse applications which see the `xterm' terminal type and
> expect it to behave like an xterm.

We are in agreement then.

> A couple of mails ago you mentioned a specification for bidi
> terminals. I've just looked back through all the mail we've
> exchanged, and I don't see a link to that spec. Can you point me to
> it, please?
> 

I think you are referring to this post:
http://www.arabeyes.org/archives/developer/2002/June/msg00020.html

All this and Kamal is still to give his take on the matter ;)

Cheers!
-- 
-------------------------------------------------------
| Mohammed Elzubeir    | Visit us at:                 |
|                      |  http://www.arabeyes.org/    |
| Arabeyes Project     | Homepage:                    |
| Unix the 'right' way |  http://fakkir.net/~elzubeir/|
-------------------------------------------------------