• From: Len Bullard <len.bullard@u...>
• To: Ken North <kennorth@s...>, xml-dev@l...
• Date: Mon, 11 Feb 2008 16:42:36 -0600

All true, Ken, but in those days (and I programmed in COBOL as my second
language), the rewards for the pain were significant and as you point out,
the limits of the hardware were such there were few commercial alternatives.

IOW, it isn't that people can't learn the syntax tools or that they aren't
supportable, it is, why do that to an innocent who has done you no harm?

And SQL injection wasn't a problem at the scale it is today.  BTW, is that
kind of problem showing up in XQuery systems too?

len


From: Ken North [mailto:kennorth@s...] 
 
Len Bullard wrote:
>> One might argue that syntax based systems, even as well-thought through
as
>> SQL, are simply not the right interface for 'non-programmers'

True, but in the 1960s that approach was dictated in part by computer
hardware.
Except for specialized applications, computers ran batch-oriented operating
systems with punched card readers as the standard i/o device. Queries were
submitted using a deck of 80-column punched cards. But that era saw a
variety of
systems that implemented a query language for non-programming users,
including
GIM (TRW) and GIS (IBM).

One of the pioneering efforts was MEDLARS (Medical Literature Analysis and
Retrieval System) at the National Library of Medicine (NLM). The first
MEDLARS
implementation in 1964 supported searching of the world's most extensive
collection of medical literature, with new publications being indexed and
abstracted every month. By 1969, NIH had won plaudits and funding for
MEDLARS
II, which included Linotron photocomposition systems. The FORMAT command
was used to format data for a phototypesetting system used for print
publications such as Quarterly Cumulative Index Medicus.

My involvement with MEDLARS II was developing the query processor that in
1970
implemented a query language for non-programmers (librarians, indexers). The
system had the capability of doing ad hoc queries and storing repetitive
queries
for better performance, such as for printing recurring publications. Our
goal
was that a user had to be no more technical than

- able to specify simple commands (SEARCH, FORMAT)
- correctly spell MeSH terms - search criteria such as 'diabetes'
- be able to check a list for terms that had been indexed for searching.

Because core memory and disks were expensive in those days, indexing was
selective. Hardware limitations dictated the use of a qualifier to indicate
a
term wasn't indexed (requiring a linear search for that term).

NLM had domain experts who controlled the vocabulary of terms used for
searching. Medical Subject Headings (MeSH) are still in use today:
http://www.nlm.nih.gov/pubs/factsheets/mesh.html

MEDLARS morphed into MEDLINE and PUBMED, with international databases and a
user
community that's grown way beyond librarians. Of course people searching
medical
literature today don't need to learn the syntax of a query language.

As for today's syntax-based systems, query tools can provide
context-sensitive
help for users composing XPath expressions or XQuery statements.


This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the sender. This message contains confidential information and is intended only for the individual named. If you are not the named addressee you should not disseminate, distribute or copy this e-mail.