Re: [Bug-gnupedia] Important decisions should not be rushed

[Bob Dodd]

--- Richard Stallman <address@hidden> wrote:
> Over the years, I've noticed a behavior pattern: when a programmer
> feels enthusiastic about a project whose goal is developing
> information (not software), he is likely to respond by saying, "First
> let's develop special tools to do this project with."  There are two
> potential problems with this pattern:
>
> 1. In many cases there are advantages to using standard software,
> and using something ideosyncratic would actually be worse.

That's very true. Of course there has to be standard software to work
from, or that can be configured to do the job. You also have to look
carefully at the licensing issues involved.

The problem I can see with "information" based projects, is that the
standard software tends to be basic relational databases (of whatever
flavour), and that software dealing with organisation of the
information, and the lifecycle of that information tends to be very
"hard-coded" to a particular view of the world. Fine if you hold that
view, but real a pain in the neck if you don't.

There is also the possibility that you may find article-management
software that exactly matches your view of the world, but is tied to an
underlying data source that is inappropriate for your use (e.g. maybe
it doesn't scale too well, can't mirror properly etc.)

That's why I think it's important to abstract the notion of
"encyclopedia" into it's component parts: data source, organisation,
and content management. It's not that you can't re-use existing
software (e.g. Nupedia's) in any of the three abstractions (or indeed
have the same software providing solutions in more than one abstraction
of the problem), you can, but it's still important to have a clear
analysis of the problem so as to fully understand the consequences
associated with particular platform/software component selections.

 
> 2. If people work on tool development, instead of the project's own
> goal, that can mean nobody works on the project itself.

That's true if you believe that everyone who offers support is
interested primarily in writing, reviewing, and organising material.
I'm not convinced that that is entirely true. Actually, you only need
to look at the postings to see that even people who would love to
contribute entries to the project are also passionate about how
information is organised and retrieved. Removing that aspect of the
project doesn't get you any more articles, just an intellectually
stunted view of information content.

That said, clearly if the emphasis became wholly tool oriented, we
wouldn't get our encyclopedia (though I guess it's possible to argue
that you may get many more, and that the general quality of free
encyclopedias would improve). What we need is a balance, which I'm sure
is what you're arguing for, but that balance needs to take into account
all the knowledge, skills, and interests of the volunteers who will
build and maintain the project. Without the volunteers, we wouldn't
have a project.
 
> Every project that asks for support in the community of programmers
> but whose focus is not software development has to guard against
> getting sidetracked into development of tools.
> 
> There is an additional problem with trying to develop tools too
> early:
> it's hard to tell what kind of tool is useful until you have real
> experience doing the project.  This means that it is premature to
> work
> on any sort of support platform until there is a considerable number
> of articles and a number of people writing articles.  Imagine trying
> to design programming tools as the first program on the first
> computer.  How would you know what they should do?

I'm sure you don't mean what you just said... Libraries and
encyclopedias have been around for thousands of years. You can even get
degrees in librarianship. To say that the organisation of encyclopedia
entries is an unknown, or that the aspects of how to classify content
have not been discussed to death, is just silly.  You may as well say
that my company doesn't need a customer billing database until we've
sent out a few bills and seen what responses we get.

If you don't select (at least a basic) information model to organise
the entries you receive, that will allow you to search & categorise,
how will you know what you've got? You will also need some sort of
annotation for them so that you can record their status. Once you have
a model, you need tools to make the model work.  I don't say they have
to be the final ones (and almost certainly wouldn't be), but they have
to be good enough. That requires software analysis, design, and
implementation (which in itself requires some platform issues to be
handled). The alternative I guess, is a pile of paper on Hector's desk
and a rolodex.
 
> Once encyclopedia work reaches the point where there is enough
> experience, then it will be useful for people to look at how work is
> done, and figure out what sort of tools could make it easier.
> 
> We're talking now about merging GNUPedia and Nupedia (which would
> switch to the GFDL as license).  This would mean we'd have a lot more
> articles; perhaps the combined experience of the two projects would
> be
> enough to enable people to see sort of tools would be useful.
> 
> 
> 
> 
> 
> _______________________________________________
> Bug-gnupedia mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/bug-gnupedia


__________________________________________________
Do You Yahoo!?
Yahoo! Auctions - Buy the things you want at great prices. 
http://auctions.yahoo.com/