-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

I think Reza's points are well taken indeed, particularly with regard
to the regex-thing. (for 1 thing : regex is regularly not very 'portable')

It is what I call the "TwoStepBuilder"-issue.

Reza's approach in creating a *fast* workable data set from the XML
resources goes much further then the derived sets my OpenDDR-fork
uses, but it too tries to cut out (or down) the 'builder' question as
much as possible. Because indeed a lot of expensive regex-jazz goes
into identifying the right builder but once there the internals,
irrespective of the builder, are, generally speaking, identical. Going
straight at it, cutting out the builder-identifying-steps, makes much
more sense.

I do not know yet, after incorporating Reza's latest updates in the
.Net version I constructed here, whether over a larger test data set
the current method, using the current XML resources' format will work
'100%'.
I seem to recall I tried something similar over a year back and ran
into trouble : a device with Id 'MB870' springs to mind (this device
is no longer in the current XML resources, at least not with that Id).

I will make both the .Net version of Reza's java client and the test
results over the big data set available ASAP in the course of next week.

In general I very much agree with Reza's proposed logic, but as he, I
think implies, we seem to be tied to the XML resources' source and
their format.

esjr
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.8 (MingW32)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQEcBAEBAgAGBQJRx4nqAAoJEOxywXcFLKYcHysIALCotX8sT8R92RTLsN88d+50
Uz32H/GorTOmyppF9X7iUihO1n5+RwnXGeltfmaLCYgdeIxUfS05R4x39cuo6400
UhCwolxljx6wl13VSh5JoTmg1ZG6LuapLrr8ZHf7TA4kzGSQPmP0ZTWPcOgesDro
pEcr94xaxYxmW9g3xDL+62CLeJqq+uG7YSsKj51P5rYfOYDocIykr29QkYEdmPci
ni91dCi5PXzOFTt0PKddeVg3VJjnRfevQVb+yMZoWE5AIzzjwRUXDsSeFBch/iOC
b40GuZHdl7V5Oondrh+uBn5knTRersSbsSCSA0N0Rwxiinkc2P8pjtVbSLmNOT0=
=BwvZ
-----END PGP SIGNATURE-----

Reply via email to