"Andrew C. Greenberg" wrote:

> >Hi people,
> >as I was listening to you, while you were speaking about Squeak
> >documentation and Squeak books, I said to myself: "why not to tell them
> >what I'm missing in Squeak". So...
> >I am total newbie in Squeak. Before I became Squeak fan I worked mostly
> >with M$ products, like M$ Visual Basic and M$ Visual Studio, so I'm maybe
> >very lazy user (in searching "how to do something" things). I like index
> >of topics which is used in help files in Wins. When I want to find how
> >to do something, first I'm searching for such topic in that list. In 80% I
> >find what I want very quickly. This is what I'm missing in the Squeak.
> >
> >I think taht source code in Squeak is okay, in my opinion it's not really
> >necessary to coment all places in Squeak sources. For me it seems to be
> >"self commenting" code. But I lose too much time with searching (and
> >guessing, where it can be) for some way to do what I want. Reading of some
> >book is for me (like for many others) the last way, after few hours spent
> >with endless searching for some topic.
> >"What class I have to use when I want to do this?",
> >"Is this the best way to that?",

At this point, collecting an interesting set of <this>s and <that>s would be
helpful.
What specific higher level questions need answering?

> > I think that
> >surfing in sources is waste of time, isn't it?

Not if they can be tagged to match your questions.

> This is interesting, and should be fairly straightforward once the
> Squeak image has, at least, solid class descriptions for all
> important classes.  It should be a simple matter to produce a
> first-cut inverted index and browser/search tool for finding relevant
> classes -- indeed, one could be built that would constantly update
> itself as the image comments change.

It'll work much better if there's a common understanding of the
questions the index should help answer. That'd help comment writers
use matching phrases.

-dms