Kenneth Whistler wrote:
> Of course, like nearly every online document project, this raises
> some questions:
> A. What EXACTLY is the standard? When I look at a document in the
> online repository, what guarantee do I have that it is identical
> to a published standard, if the standard itself is a registered
> version of a hardcopy document?
First of all: Even with paper copies, you are not sure what exactly
the standard is. Paper copies often contain typographical errors which
affect the interoperability of products designed according to this
standard. It usually takes over two years to get a technical
corrigendum published, even if the flaw is obvious and discovered
three days after standard has gone into print. This is not some
academic theory, it happens in real live: Such serious typographical
errors are e.g. in POSIX.1b (IEEE Std 1003.1b) or in the JBIG
specification (ISO 11544); I have implemented both.
I have found similar silly problems in almost any ISO standard which I
have read so far (e.g. ISO 8601 said before the first corrigendum that
an hour has 61 minutes and in an ISO 8802 copy which we have bought
from ISO for a lot of money a few years ago, several pages have simply
been missing). In the case of POSIX.1b, the buggy version has already
been implemented widely and a technical corregendum about which I know
is already under preparation and will be published in 12 months. It is
a complete mess and even although I have full access to the paper
versions of all relevant documents, I do not know now what I am
supposed to implement. As electronic standards can be fixed and
redistributed much easier and faster than widely distributed paper
versions, such problems would be much less serious. Problems like
differences between the final copy corrected by the committee and the
printed version and missing pages are much less likely with electronic
document distribution.
In addition, you can of course easily add some disclaimer to the
electronic version that says:
"Only the paper version obtained directly from ISO or its member
bodies represents the authentic official standard. The electronic
version is only a best-effort representation of the content of the
original paper version. ISO is in no way responsible for any
inaccuracies in the electronic representation."
This will guarantee ISO's revenue, because implementors will still be
forced to buy the paper version, but it will give studentes and
researchers, which can not affort to buy holly ISO paper anyway,
browsing access to the standards in order to let them determine
whether these documents are relevant for them at all. Most likely,
this approach will even increase ISO's revenue, because MUCH more
people can read the standard before they buy it and they are therefore
more willing to spend some money for the official paper document,
because now they will know from the on-line browsing that this text is
really relevant for their work.
In the long-term, it would be better to make the electronic version
the official one, because in contrast to paper copies, electronic
files can very easily be authenticated using cryptographic message
digest algorithms (e.g. using the SHA standard described in FIPS PUB
180-1 or the MD5 algorithm specified in RFC 1321). This way, the
autenticity of the document can be verified by anyone with on-line
access within seconds. (Check the PGP public-domain software if you
are not familiar with public key cryptography and message
authentication.)
Question B should also have been answered by this.
> C. If this online repository is implemented as a Web repository, I
> presume it will be posted in HTML and that pages will start
> consisting of cross-linked hypertext. How do I know that I have
> reached the end of any particular document or standard once the
> cross-links start going up?
This is simply achieved by introducing a URL naming convention. As
long as the URL has the form
http://www.iso.ch/standards/en/ISO-11544:1993/....
you know that you are inside ISO 11544:1993 and not in any other
document. In addition to the HTML version, you could also offer a
PostScript version of the document which contains exactly the contents
of one ISO document. PostScript ensures that the page layout and
content is exactly identical on all printers (which HTML does not) and
since the free availability of the ghostscript software, everyone can
print PostScript files easily on virtually any printer.
> Who is signing up to maintenance of
> the links and as the standards czar to define the hypertext
> topology of what gets posted?
As far as I know, ISO standards are produced anyway using a SGML
master version of the standards text. If ISO's SGML document type
definition for standards has been carefully designed, it contains
anyway a little bit more information than what is visible. E.g. in the
reference section of an ISO document, the SGML version should contain
tags that mark the exact ISO designatior of other referenced
standards. From this master version of the standard in SGML form, an
algorithm can easily construct the HTML links fully automatically.
This way, the hypertext topology is just an algorithmic transformation
of the information already marked-up in the electronic SGML source
version of the standard which ISO uses anyway already today to print
the document. I see no problem here; all this can even be implemented
completely using existing well-established ISO standards (SGML, DSSSL,
etc.).
Markus
-- Markus Kuhn, Computer Science student -- University of Erlangen, Internet Mail: <mskuhn@cip.informatik.uni-erlangen.de> - Germany WWW Home: <http://wwwcip.informatik.uni-erlangen.de/user/mskuhn>
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:30 EDT