Emacs & Hebrew

[Nadav Har'El]

On Wed, Jun 13, 2012, Eli Zaretskii wrote about "Re: Emacs & Hebrew":
> You mean, disappointing?  Yes, it is.  To hear such questions from
> Nadav Har'El, of all the people.

Sorry to disappoint you...

One of my random signatures go "better to be thought a fool, than to
open your mouth and remove all doubt". Too bad I opened my mouth :-)

> And since the bidirectional display for Emacs was developed in almost
> complete isolation from this community -- not a single input or
> response to several design discussions I posted -- why would anyone
> assume that an essentially one-man project will not end up being a
> complete disaster from usability point of view?

Because I know you, and don't believe you are capable of producing a
complete disaster? :-)

> > Bidi is great for writing texts, but since until now writing Hebrew
> > text in Emacs wasn't a great idea, people didn't do it. What they
> > did do with Emacs is writing code, editing config files, and similar
> > things. With those, Bidi is sometimes a distraction, not a desired
> > feature - so people want to be able to turn it off.
> 
> How do you know it's a distraction, if you didn't yet try it?

Let me explain what I meant, and why I wrote my question.

As you may know, Hspell has a perl script which takes Hebrew base words,
and inflects them (e.g., produces plural, possesives, etc., of nouns).
The Perl script is, of course, mostly ASCII, but with various short Hebrew
strings sprinkled in it.
It is, in my opinion, less confusing to view these strings without BIDI
applied, which is why I edit it with regular non-bidi-capable vim.

As an example of the confusion that bidi might cause me when editing the
Hspell source code, conside what it might to do a regular expression:
Imagine that I wrote s/א/ב/ with the intension of switching aleph into bet,
but (I just checked...) the bidi algorithm applied to it will display the
order *reversed* making it confusingly appear that the code is changing bet
into aleph...

I think you must agree that the unicode bidi algorithm was never designed
for such use cases - and I'm not "blaming" these issues on your
implementation, of course - and I don't even need to try your
implementation to know it will have this - or similar - issues when
editing source code. And editing mostly-ASCII source code is 99% of what I
personally use Emacs for.

Again, nothing I said or asked applies to writing *text*, such as email.
For text, bidi is definitely required, and really welcome. I sorely miss a
bidi-capable editor for my email writing - unfortunately I have a 20 year
old habit of writing emails in vi, so it'll take me a bit of time to decide
to switch to emacs also for emails. But now I know I definitely should!

-- 
Nadav Har'El                        |                 Wednesday, Jun 13 2012, 
nyh at math.technion.ac.il             |-----------------------------------------
Phone +972-523-790466, ICQ 13349191 |A Nobel Peace Prize? I would KILL for one
http://nadav.harel.org.il           |of those.