status for release 1.0

[Peter Dobcsanyi]

This is a status report regarding the release of 1.0.

    Current version: 1.0rc3      (revision 1.78)
    Release date:    end of October
    
    Settled:  indicators for 1.0

    Pending:
	numerical, exact/inexact semantics 
	moving ext-rep documentation to LaTex
	update/finish the ext-rep documentation


INDICATORS
----------

Summary of the recent (and ongoing) thread about indicators, in
particular how that effects the 1.0 release.

<partially_balanced> indicator and <partial_balance_properties> subtree
have been withdrawn from 1.0 since there are many related issues to be
resolved. Hopefully, partial balance will make it into 1.1.


The "affine" indicator has been placed into 1.0 as <affine_resolvable>.
Semantics: 'flag' is true iff the block design is resolvable and any two
blocks in different parallel classes of a resolution meet in a constant
number 'mu' of points.

The current RNC definition is:

    element affine_resolvable {
	attribute flag { "true" | "false" | "unknown" },
	attribute mu { xsd:positiveInteger } ?
    } ?

I am not sure about "unknown". Do we need it here?


The <connected> indicator has been extended to show the number of
connected components. The current RNC definition is:

    element connected {
	attribute flag          { "true" | "false" } ,
	attribute no_components { xsd:positiveInteger } ?
    } ?


NUMERICS
--------

There is work in progress regarding the numerical semantics and the
relationships between exact and inexact computations.
See the thread 'numerics' started by

    http://designtheory.org/pipermail/developers/2003-October/000182.html

There seems to be a consensus that we are going to take into account all
available information obtained by possible exact methods.

A lot need to be done here, at least in the documentation. It is quite
possible that not all aspects of this issue can be resolved for 1.0 --
for example, the formalism for indicating inconsistency.

As a minor step, the semantics and scooping rules of the attribute
"precision" was laid out in the post above. Accordingly, 'precision' was
also added to <list_of_designs> and <block_design>.


DOCUMENTATION
-------------

The maintenance and production of the ext-rep doc in the current format
consumes too much (mainly human) resources. In the same time, we cannot
produce satisfactory quality printable format. Also, based on the
web-hits statistics of the beta release, it is fair to say that people
were more interested in the PDF format than in the HTML. One reason for
that could well be that not every browsers could render equally nicely
our document.

Therefore I recommend to transform the current documentation into Latex
and produce high quality printable format and somewhat lower quality
HTML format. The extractor programs will still be used for RNC and
ext-rep segments in a similar way.

I am going to do the transformation to LaTex myself with the most recent
version and when the new framework is ready will announce it on the
list. Meanwhile, please refrain from updating the ext-rep doc.

A question to TeX experts:

    Does anybody know a way of executing a shell command from (La)Tex
    source during the typesetting?

A related issue:

    The most recent version of the Tehran conference paper with new
    style file is in the CVS. See also my earlier post

	http://designtheory.org/pipermail/developers/2003-September/000138.html

    It is very likely that I will test the new document production
    method with this article. Regardless, you can keep working on it.


Feedback please, also let me know if I left out something important.

--             ,
    Peter Dobcsanyi