Lists Home |
Date Index |
- From: "Hutchison, Nigel" <email@example.com>
- To: firstname.lastname@example.org
- Date: Fri, 16 Apr 1999 16:55:44 +0200
Title: RE: ASN.1
From: Mike Spreitzer [SMTP:email@example.com]
Sent: Friday, April 16, 1999 3:31 PM
To: 'Hutchison, Nigel'; firstname.lastname@example.org
Subject: RE: ASN.1
Nigel Hutchison wrote: [[
I would have thought that the best way of dealing with this issue is to use
a "pleasant" syntax which was easy to process and implement another layer
to compress the payload for transmission.
Best in some contexts, but not all. The compression layer has runtime
costs, in both code and memory footprint, and processing time. In some
contexts (e.g., very resource-constrained items like cell phones and Palm
Pilots), these costs can be significant.
[Nigel Hutchison] I had some experience recently in trying to devise a RPC XML DTD which was nice and compact. So I did also sorts of tricks with short tag names, repetition conventions etc etc. This had the effect of iincreasing the creation and parsing code significantly. I imagine that there is a considerable footprint and CPU penalty is this optimisation. II also found the prospect of testing and debugging this quite daunting. I then realised that when I send and receive XML documents via my analogue modem they are effectively compressed and decompressed by the hardware - so I was wasting my time (in that scenario at least). I would have thought that all cellphones would do compression and decompression when they send and receive data - is that not so?
I also found out that Mainframes have firmware supporting LZ compression these days.