World: r3wp

hmm, not sure .... is difference any different from date1 - date2

I mean - when substracting two 'now values, the time is not being 
taken into account ..

>> a: now - 125:00:00
== 19-Mar-2006/9:08:07
>> a - now
== -5    ;; five days
>> difference a  now
== -125:00:16    ;;; 125 hours
That's the difference!
>>

There must be no timezone present

why should be difference call different to simple usage of minus 
op?

it's a strange hack -- difference was originally for sets. So making 
it work for time is seriously non-orthogonal.

I hate those hacks

Maybe we need three types instead of two. date! time! and idate! 
or stamp!

difference hadles timezones too.

I don't like hacks, but the minus op is a very practical shortcut. 
Should it try to be smart and return days if no time exists for either 
arg; otherwise do the same thing DIFFERENCE does (and what time do 
you assume if only given for one arg; midnight UTC?)?

Can rebol do hashes of hashes, arrays of hashes, hashes of arrays, 
data structures within data structures, similar to the way Perl does 
it with references?

I haven't actually seen documentation which explains concisely how 
to accomplish this with Rebol

it would be interesting if it could also do arrays of objects, and 
hashes of objects

but let's say I set some words to some blocks, and I put those words 
into another block, do I always have to reduce to actually see the 
stuff pointed to by the words in the blocks, or is there a way to 
actually insert the content of the words themselves into the blocks

You can reduce or not; that is, the choice of when to evaluate is 
up to you; and, yes, any type of block can contain any type of data 
nested to any level.

You need to be aware that the behavior of lists is *slightly* different 
than for blocks and hashes, WRT what INSERT does, etc. The Core manual 
covers it correctly I think.

hmm

but I mean, I don't want to have to reduce, I want to insert the 
actual values of the words into the list, not the word itself

when I insert the word, I have to reduce to show what the contents 
of the word actually are

let's say a: [value1 value2 value3]

and I want b to have a in it

b: [a] will have the word a, not value1 value2 value3

what I do b: [a] and have it put value1 value2 value3 in b instead 
of a itself

you can do B: REDUCE[A], but if don't want to reduce you can also 
do sth. like B: [ ] APPEND B A
But I think what your looking for is sth. like
	>> a: [1 2 3] b: [ ]
	>> append/only b a
	>> probe b
	>> [ [ 1 2 3 ] ]
	>> clear a
	== [ ]
	>> probe b
	== [ [ ] ]
	>> append a 4 probe b
	== [ [ 4 ] ]

why exactly is it called reduce also?

Because it reduce the words to their values?

It reduces the expressions found within the block (and it returns 
the results in a new block).

I am also perplexed why rebol doesn't have a substring function, 
sure even if you can create one with a few extra commands, it seems 
to me like most people don't want to have to invent a language, just 
to use the functionality that exists.

or I should rephrase that, to use functionality that exists in other 
languages

I hope rebol 3.0 has a substring function built in

even if source substring shows me the copy/part at stuff

A line in the sand has to be drawn somewhere.

substring is far more intuitive for someone who hasn't used rebol 
before than copy/part at

Gregg has posted a dialected substring function.

rebol's help also should have a see also at the bottom of commands, 
to direct attention to other rebol commands and functions of interest

the language as a whole feels very fragmented, I recognize the power 
of Rebol only because of its lispy nature, but I feel limited compared 
to a language like perl, because examples are far too scattered about, 
and the names for functions are often oddly named or sometimes the 
functionality exists but is so obscured that I never knew it was 
there at all in the first place

I hope more is done to actually document the language when 3.0 rolls 
out

and sure perl ain't the prettiest languages on earth, but it is well 
documented

for example, how could I take a date, and add 3 months or 3 weeks 
to the date?

a CPAN for Rebol?

hmm, that is one idea, hehe

But I am actually talking thorough discussion and documentation of 
every nook and cranny and how to do basic things people want and 
need to do with the language

something along the lines of the rebol core manual, but more in depth

having to re-invent the functionality that other languages just give 
you is a serious pitfall if you ask me

that was one of the things that drove me crazy about lisp, yeah it 
is a powerful language, and yeah it is a programmable language, but 
to some degree the only way to do anything really useful in the langauge 
was to in essence program lisp to function similar to other languages 
you were already familiar with, or to in essence extend lisp with 
functionality it should have already had before you even started 
programming in it it

in some ways this might be considered an elitist atttitude on the 
part of the language writers.

If it's not there, write it yourself.

Better to have a way to point to mezzanines that implement the common 
functionality that people want.

I dunno, I think that languages like java or .NET would be far less 
successful if people had to implement a majority of the class library 
by hand, anytime they wanted to take advantage of advanced functionality. 
There is just some functionality that people come to expect regardless 
of what higher level language they are coding in. And for a language 
that supposedly can be used to parse and manipulate text and data, 
not having a built in substring function is pretty crazy

Rambo your suggestion.