Why is Elisp slow? (was: Re: Naming, and Gnus functions length [Was: Re: Knowing where a function has been used (bis) [Was: Re: Optimising Elisp code]])

[Emanuel Berg]

Eli Zaretskii wrote:

>> I once said to the Gnus developers, why
>> don't you break down those age-long
>> functions, to make it much easier to read
>> and maintain the code?
>> 
>> They said Gnus is slow, Elisp is slow, and
>> breaking the defuns up into neat modules
>> would make it even slower.
>
> They are right.

Why is Elisp slow?

Is it funcall overhead that is slow in
particular or is Elisp slow in other aspects
as well?

Is Lisp slow in general? For example, compared
to C? If so, what is the reason for this?

How much faster does it (Elisp) get with the
byte compiler?

How much faster can one get it with inlining,
assuming it is done by a skilled hand?

Are there other speed-up methods other than
inlining. (I mean specifically, not "write
better code".)

Is there like a list of things, functions
perhaps, or data structures, that shouldn't be
used unless necessary, because they are slow?

Is there a tool that I can enable, then use
Emacs a whole day (or night rather), and then
the tool tells me "everything is OK, except for
when you do this, that sticks out as something
that drags down the tempo, so maybe you should
have a look at it"?

Facts for fans: This post contains 12 "?"s!

-- 
underground experts united
http://user.it.uu.se/~embe8573