At 10:46 AM 3/4/96 CST, Wendell Piez wrote:
>As for other pc-based tools, I don't have broad experience but haven't
>come across anything to recommend. I do hope other readers of the list
>post their observations to this thread.
A few comments about WP SGML. Our project has been using the beta
version of WP SGML in earnest since the beginning of January. Most of
our users are familiar with WP but novices to the world of SGML;
consequently, their views on the process of tagging information are very
interesting. Our project is academic in nature; we are developing an
integrated history of women's writing in the British Isles. We are
currently capturing information with locally developed dtds.
By and large we are pleased with WP SGML, but there are a few quirks
that we hope disappear when a commercial version of the software is
released. The most serious of these quirks is WP SGML's tendency to
cause keyboard freezing. Almost all of our computers have experienced
it -- our Windows '95 computer more-so than our Windows 3.11 computers.
When the keyboard freezes, nothing will work but a hard boot.
Fortunately, WP's back-up mechanisms mean that we have not lost much
work. Since most of our taggers are part-time graduate research
assistants (who often work evening hours when computing support folk are
not around), the stress levels induced by the keyboard freezing are not
negligibleu1 Has anyone else experienced this?
1. Ideally, we want our taggers to work exclusively in SGML, validating
their documents each time they save. By doing this, we hope to avoid the
risk of WP command-related errors creeping into our documents.
Furthermore, working in SGML exclusively keeps taggers on their toes,
because they know they have to validate their documents before they
leave them. WP SGML makes it difficult to work exclusively in SGML.
For example, if a tagger must exit a document before it is valid, WP
SGML warns that the document is invalid but will save it as SGML anyway.
When the document is opened again, however, WP SGML has trouble
interpreting any invalid information. The software will take all
invalid information and move it outside the element that it feels the
information should not be residing in. For our project, this means that
a simple error in the ordering of elements can result in WP SGML dumping
3/4 of a document outside the document end tag. Because we are using
brand-new locally developed dtds, we are constantly needing to tweak our
dtds to better suit the project's needs. Whenever I introduce even a
small change to a dtd, however, I have to warn all taggers that I have
"broken" their documents.
Because of these problems, I have started asking taggers to save
their documents as WP instead of SGML. This way, changes to the dtd
result in invalid documents instead of broken ones. I feel most uneasy
about having to make this compromise though.
2. Style sheets: When a style sheet is created in layout designer, it
is affiliated (by file name) with a particular dtd. Because our dtds go
through multiple versions during their creation, we have to either a)
start from scatch to create a style sheet for each version of the dtd (a
cumbersome process given the breadth of our dtds) or b) have a standard
name for the latest version of any given dtd in order to keep one
standard style sheet, and rename archival versions of the dtd. We are
doing the latter, but every now and then we want to call up documents
that were created in older versions of the dtd--mainly to see them
before the new dtd will break them. The archival names of the dtd then
have no style sheet.
3. Viewing attributes: We capture a lot of information in attributes.
WP SGML will not let a tagger view attributes other than in a
frustrating cursor-sensitive reveal codes screen. When we check
documents, often we will print things out in ASCII versions of the file
just so we can see attributes. Doing this, though, means that we loose
all style sheet formatting.
Generally, we are pleased with WP SGML. It has been remarkably
staight-forward for our research assistants to learn and to use adeptly.
The complaints listed above are meant to let others know what some of
the software's drawbacks might be.
An Integrated History of Women's Writing in the British Isles
3-5 Humanities Centre
University of Alberta
[log in to unmask]
(403) 492-8142 (fax)