Re: [Lua-users-stdlib] Upheaval (blame Jamie Webb!)

[Johann H. <jhi...@ya...>]

On Feb 7, 2004, at 6:15 PM, Jamie Webb wrote:
> On Friday 06 February 2004 03:42, Johann Hibschman wrote:
>> Well, "list" is a natural language word, so it's broad and doesn't
>> promise much.  A "vector" is a rather specialized technical word, more
>> specialized than even "array", so it promises a lot.
>
> That's just it though: all three are rather specialised technical 
> words. The
> exact way in which they are specialised depends on the field you work 
> in.

Well, there is still a difference.  "list" is common language, that's 
been co-opted by many programming languages to mean linked lists, but 
it's certainly not omni-present.

For example, I do a lot of Python, and Python lists are vector-like 
lists, not linked lists.  However, it's certainly true that the 
list/vector difference is maintained in Scheme, Common Lisp, Java, 
OCaml, and probably many other languages.

It doesn't really matter that much.  I can always rename any "vector" 
module to "list" and be done.  I'm just lazy, find it easier to type 
"list", and think its natural-language overtones mesh better with the 
word "table".

>  Since Lua is a mixed imperative/functional language, we should take 
> our
> definitions from those worlds: Arrays are fixed-size, n-dimensional 
> mutable
> structures with O(1) random access, O(n) inserts, etc., Vectors are
> single-dimensional, mutable and resizable in imperative languages, 
> with O(1)
> random access, O(n) inserts, etc. Lists are mutable and resizable in
> imperative languages, and have O(n) random access and O(1) inserts. 
> Data
> Structures books may qualify such lists as linked lists, but 
> programming
> languages such as C++, ML, and LISP don't.

Actually, this reminds me.  What are the times on Lua tables?  I know 
that integer indices are special, but do they get true O(1) access?  
Presumably, other keys are O(log n), as usual.

> Also, the implications of a misunderstanding differ: If a user were to 
> assume
> he was dealing with a mathematical vector when in fact he wasn't, that 
> could
> only be as a consequence of a more direct misunderstanding of the 
> operations
> he is performing, and the interpreter is likely to complain very 
> quickly.
> OTOH, if a programmer were to assume he was dealing with a linked 
> list, he
> could quite happily produce a working program, only to discover once 
> he tries
> it on production-scale data that it runs N times slower than expected.

That's certainly true, but I think it's very unlikely.

>> At work, I'm converting over some old APL programs, so I'll probably
>> have some extremely screwy ideas in a few weeks, if I can ever get the
>> blasted font to work.  Now _there_'s a concise language.
>
> Mmm... I thought those people who still dealt with APL tended to use 
> ASCII
> representations these days?

Nah, ASCII representations are for the weak.  You need the funny arrows 
and things.  :-)

In any case, I don't see much of a point in continuing to discuss this, 
since I can always rename packages to whatever I prefer.  If you would 
really rather call it a vector package, we can just go with that.

Cheers,

Johann