Print

Print


BACKGROUND (sorry for the detail, but it will help explain the questions
at the end):
 
For my text editing project I have built a PARADOX relational database
over a number of years. This database contains brief entries for 111
Pirqe R. El.  manuscripts and fragments. The PARADOX database (mss.db)
is a composed of 111 ROWs with 14 COLUMNS each. (This is, as I
understand it, the TEI equivalent of 111 ROWs with exactly 14 CELLs in
each). Three of the columns contain as text a short identifier which  is
used to link to other databases which contain more information on the
same listing. Thus, in mss.db column Library, one might find the number
<>25</> in a number of rows containing references to Geniza fragments.
This ID number links to a lib.db in which <>25</> is the unique
identifier for the Cambridge Library. In creating a report in PARADOX,
the two databases can be accessed, and instead of seeing <>25</>,
one could tell the report to show the cell which contains <>Cambridge</>.
 
In addition, there is a partially hidden column for extended comments,
the equivalent of a NOTE with extended text, describing the manuscript
in many of the entries.
 
I am in the process of converting this PARADOX database to TEILITE. The
purposes of the conversion are:
 
        1) to make the database available on the WWW.
 
        2) to put the database into a form which can be linked with each
file containing the transcription of a ms., so that a reader can quickly
get to information on the ms. from the transcription through an XREF.
 
There is no problem in exporting the database(s) from PARADOX in text
format. Each ROW is provided with a line break, each CELL entry is
marked with "" and separated from the next with a comma. Using a simple
text search and replace one can turn all <>",</> into </CELL> and the
remaining <>"<> into <CELL>, quickly add <ROW with an ID> and </ROW> and
insert into <TABLE>.
 
But that's where the problems begin. For example: With the quick
conversion outlined above, the CELL tags presently have no attributes,
and thus are undifferentiated. With some more effort, I could export
each PARADOX COLUMN separately and put a ROLE attribute designation in
the CELLs to indicate what the CELL contains.
 
QUESTIONS:
 
1) Is there any quicker way to deal with this? I'm looking for
pre-existing magic, which would put an abbreviation of the COLUMN name
in the ROLE attribute of each CELL in that column, and do it for all
columns automatically.
 
2) What tools does one use to query a TABLE in a SGML/TEI database? Such
search mechanisms are built in to PARADOX and other commercial
databases, but how does one get information from an SGML/TEI file? For
example: list the "catalog num." "microfilm num." "chapters" of all
<>f</> (=fragments) in "manuscript type" found in "library" <>25,</>
(=Cambridge)? In PARADOX, one would check a box listing the column title
(here given in "" marks).
 
        a) what platform?
 
        b) what software does one use, or are we talking about programming
language, and if so which?
 
        c) can queries be created for information found in tables
contained in different .sgml files?
 
3) Finally, is what I have in fact best represented by a TABLE, or
should some other catagory such as LIST or BIBL be used?
 
Sorry if these questions are too simple. Any responses would be greatly
appreciated. Thanks in advance.
 
Lewis M. Barth
*********************************************************************
Lewis M. Barth                  Office Phone: (213) 749-3424
Hebrew Union College            Office FAX:   (213) 749-1192
3077 University Avenue          Internet:     [log in to unmask]
Los Angeles, CA 90007           CompuServe:   71231,331
                                WWW:          http://bcf.usc.edu/~lbarth/
*********************************************************************