[OT] Emacs, Eclipse, Leo (was Re: IDE

Ville Vainio ville at spammers.com
Thu Jul 15 09:15:57 CEST 2004

>>>>> "Peter" == Peter Milliken <peterm at resmed.com.au> writes:

    Peter> Let's not do the war thing :-) But perhaps an exchange of
    Peter> information?

    Peter> Me, I use Emacs - have done so for 20 years now with never
    Peter> a regret. When I

I use Emacs as my main editor/mail/news too. Too bad Emacs seems to
lack the development energy these days to match the features of modern
IDEs. I'm mostly talking about features like robust code completion
for C++ - cedet doesn't cut it yet (though it'll hit 1.0 RSN), and
there just isn't enough developers to make it work hitchlessly. There
seems to be a resource shortage on xemacs side, and GNU emacs is,
well, too RMSic to do anything modern. Now they are integrating GTK,
which is a good sign of course ;-).

Lately I've been starting to hope that Eclipse project would take the
role of Ultimate Editor That Will Be Around Tomorrow - something which
you can expect to use 10 years from now. It has its shortcomings (big,
slow, Java), but it's the best we've got so far.

    Peter> key sequences - but I don't like using mice and menus -
    Peter> takes longer than

Yes, complete keyboard is a must for something you plan on using all
the time.

    Peter> their personal use? For example the following comment is
    Peter> made in another (later) thread:

    Peter> "augment Leo's mind blowing qualities" - which means what?
    Peter> :-) Other than the fact the writer likes Leo? :-)

As the author of the comment, I think I need to explain a bit. Leo is
mindblowing, not as an editor but as an intelligence management
platform. I could easily imagine project managers using it to organize
things. It seems to have endless usage possibilities, of which
programming (or text editing is only a fraction). Ed was talking about
implementing a multi-user editor, and I imagined it would be quite
sweet to have outlines there too, from team collaboration perspective.

    Peter> code templates for where the coder must supply the
    Peter> arguments. But I guess I am not big on "syntax completion"
    Peter> - I haven't really used to the code since I wrote it,
    Peter> although I do use ELSE itself extensively in writing my
    Peter> Python code :-)

Syntax completion (or semantic completion, or "intellisense" - syntax
completion suggests insertion of code templates, which are simple in
python anyway) is very handy in multi-person projects, or projects
that have craploads of code written by other people - sdk developers,
various subcontractors, whatever. Maybe less so in Python - for some
reason Python code is usually more "direct" and elegant - but it would
help a lot. We need some kind of static type information to provide
intellisense for Python, though. It's just something that needs to be
standardized in the language level, so tool writers can proceed with
implementing the intellisense.

    Peter> 4. User extendable using either Lisp or Python ( :-) ) -
    Peter> (see the reference

Here I hope eclipse + jython will turn out to be the "preferred" way
to extend the platform. Maybe someday :-).

    Peter> 8. "Free" - in this day of powerful editors that are
    Peter> available for zero cost, why purchase one? Either shareware
    Peter> or commercial?

Yes, this is a must. Free beer & speech. Eclipse seems to combine the
best of both worlds here - the platform is free, but companies can
sell plugins for money. That way you are not "trapped" and can hold on
to your editing habits, while using some commercial stuff for exta
functionality; it won't hurt if it's company money :-).

Ville Vainio   http://tinyurl.com/2prnb

More information about the Python-list mailing list