Re: [Bug-gnupedia] Nupedia

[Thomas E. Vaughan]

On Thu, Jan 18, 2001 at 12:43:41PM -0600, Jimmy Wales wrote:
> 
> Toward this end, we have thought long and hard about what the canonical
> form of the data should be.  Hector's idea (expressed on the new gnupedia
> website) of using HTML is a singularly _bad_ idea.  HTML is a confusing
> mishmash of structure and presentation.  What we want to do is _not_
> control how people might present the data.

OK.  So you have an XML DTD just for encyclopedia articles.

> One idea I have for the encyclopedia is for competing publishers to pick
> up the data and use it to print very cheap encyclopedias for sale and
> distribution to 3rd world countries.  These countries will not be able to
> afford computers and internet access to the home for quite some time now.
> But I would love to see every household in the world have an inexpensive
> (for the competitive cost of printing, which is quite low compared to the
> cost of Britannica for example) encyclopedia _in their native language_.
> 
> Nupedia can make that happen, but only if we aren't narrow-minded about
> our data format.  I shudder to think about the difficulties of going from
> HTML to print.

Then you will need, among other pieces of conversion software, both

(1) software to translate from LaTeX to the XML, so that an author can
    write mathematical content in LaTeX, and

(2) software to translate from the XML to LaTeX, so that

    (a) a high-quality printed version is obtainable, even for math
        content, and

    (b) a PDF version of the article is available on-line for high-quality
        browsing of math content, even in the absence of MathML-enable
        browsers.

> Additionally, XML lets us do (in the distant future) some really
> interesting and clever things with a search engine.  "Give me a list
> of all the Presidents of the United States born before 1914."  This
> is possible if all the of President biographies are marked up with
> an appropriate amount of XML.  Coool!

Actually, this would be possible even if only article metadata were marked
up in an appropriate XML.

-- 
Thomas E. Vaughan <address@hidden>
CIMMS/NSSL, Norman, OK, USA