parallel lineal numerical string comparison algorithm

[john roberts]

on 5/19/03 2:09 PM, Mark Brownell at gizmotron at earthlink.net wrote:

> There would not be an advantage, SQL is already the best way. I was
> talking about the easy to use Query Language that would be for those
> that don't want to learn SQL. In a small enough database parsing the
> XML would be fast enough with PNLP. I guess it would still reguire an
> easy to use Query Language to use this as well.

XML is a great mechanism for manipulating data that needs to be displayed in
multiple environments. It offers some promise for diagnostics in certain
limited situations. IMHO it is a terrible way to store data for a database
although there my be some cases where it's useful to merge the data and the
display method.

The strengths of SQL lie in the fact that it works well and most people
working with databases are already familiar with it. However, it is not
always the best way to go.

When I started working with Valentina (V4RB) I worked with both the
baseObject methods and SQL. The SQL methods were the easiest to explain to
others; however, the baseObject methods are much faster (at least in my
timings) and much more powerful.

I think the (simple ?) Query language that Ruslan is proposing has two very
likely outcomes: 1) a Valentina/SQL speedup and (2) doing finds using much
less memory.

John Roberts