Re: browsing for past advice

[WendyPetti@xxxx]

(Note: I'm cross-posting messages from Jeff, Ray and myself to MWcybercourse 
and LogoForum. Jeff's, the first, is at the bottom of this message.)

> I agree with all that Jeff has written [see below]. (The example sited was 
very good) 
> There is one additional challenge: even those of us who have "done it 
before"
> very often forget. It is then desirable to have a archive-help mechanism 
to 
> browse and find the problem area that we are focusing on. Maybe a form of "
> browser of the archive". Cheers,
> Ray Catzel

Let us remember that the cybercourse will be on a different website and will 
not be the archive of a group. I do agree that a method for searching an 
archive would be wonderful. But it is easy to add a search engine to a 
website. My favorite search engine (which can be added to any site) is 
Google. It does a great job of dealing with phrases. I've put Google search 
on my Math Cats site to help people locate specific information. It does 
take a few weeks for new content to get indexed by Google (automatically). 
But if we do set up an online course, then even years from now, someone could 
type in, for instance "question and answer," and Jeff's contribution (which 
will be incorporated into the cybercourse) would be one of the page links 
which would pop up, along with a link to every other page in the cybercourse 
dealing with "question" and "answer."

In fact, even to this day, when I want to find an item from the old Logo-L 
from years ago, I just go over to google.com and type in my question (or 
relevant keywords), and almost always, at least one relevant item from Logo-L 
pops up!

And when I want to locate my own old Logo-L posts, or the replies to my 
questions, I just go to google.com and type in my own name, and lo and 
behold, there are my old Logo-L postings... and links to the replies!

Since Frank has raised the valid point that some discussion which would be of 
interest to the LogoForum in general might drift over into the MWcybercourse 
and thereby deprive the LogoForum, I am going to take the liberty of 
cross-posting these posts to the LogoForum; Jeff's original post is below.

Wendy Petti

Here is Jeff Richardson's post from earlier today:

Hello
I've just subscribed after reading the news on logo-l
Reading the brief on the webpage, I thought of another thing we might do;
build an archive of undocumented or hard-to-find-or-figure-out traps for
young(and old!) players. Here's an example that I struk in the past 24
hours:
I got an eMail from an upper primary school child. He was building a MW
project using QUESTION. The idea being to have something approaching
natural language processing....
He struck what seemed like a bug in MW.
When the ANSWER to a QUESTION is more than one word, like this:
QUESTION[who are you] ( and you type in ' i don't know ')
ANSWER will return:
i don't know what to do with i dont know
Note that it is NOT a List.
When you try to use this ANSWER in an IF statement, like:
IF ANSWER = [i don't know]
...you'll get a FALSE returned.
But if you know enough to think of using LIST or SENTENCE to try to
rectify this, say:
IF (SE ANSWER) = [i don't know]
...you'll still get a FALSE, yet (SE ANSWER) will return:
i don't know what to do with [i don't know]

I eMailed the ever helpful Shawn Jesty at info@xxxx

The problem is that ANSWER is a character string, not a bunch of WORDs as
it might appear..the spaces are characters.
The solution is to use PARSE. PARSE recognises the spaces and uses them as
separators to turn a character string into a List. So:
PARSE ANSWER = [i don't know]
returns TRUE.

The reason I tell this story in detail is that this sort of thing often
happens when pushing into the upper realms of project construction, and
the documentation can sometimes lack the detail to provide a solution.

As Brian Harvey says, "the difference between an expert programmer and a
non-expert is that the expert has seen everything before". Once you've
encountered and solved these problems, whether they're peculiarities of
the MW implementation, or algorithmic solutions(like getting a turtle to
draw an ellipse), you'' know what to do next time. And how to help someone
else who encounters the same problem...chances are tho' that nearly
everyone doesn't solve these things for themselves, they're helped by
someone else. Someone who has seen it before.

I hope this archive can enhance that cultural process.

cheers

Jeff Richardson