Scoble on Binary XML

Rob posts on binary XML, suggesting that it's a step backwards - as readability is a great part of XML text:
Binary folks will point out to me that binary is smaller and faster. I say "so?" Assembly is smaller and faster than C#, but does that really matter anymore except in really weird situations? and Let's just stick with systems that humans have a chance of understanding.

Leaving aside the fact that some of us do (or did, in my case) work in weird situations, there are classes of endpoints that are heavily constrained by available bandwidth. I don't think anyone is suggesting that XML exposed to users should go back to binary encoding, but small devices connected via anemic wireless can leverage binary encoding to good effect.

On Redmond

Went up to Redmond last week for the Book Publishing Partners Summit. I missed blog/group dinners with Scoble twice - there always seems to be too much to do in Redmond, and not enough time to do it. Saw some new stuff that I can't talk about, but suffice to say that missing the PDC would be a big mistake. There's enough Avalon, Yukon, Longhorn, Indigo, and Visual Studio goodness to fully occupy 3 or 4 weeks of conferencing, so I'm planning on buying tapes/DVDs. Avoid knowledge of the future at your peril.

I did get a chance to catch up with Bertrand Meyer - he was at the Rotor conference presenting work on declarative Eiffel concurrency (SCOOP) that ETH has been doing. All Rotor conference attendees had their country of origin indicated on their badges. For some reason, the ETH badges were labeled with Swaziland instead of Switzerland, putting Africa in the center of new developments in component-centered software. At least for a few days.

So what was more enlightening - the book summit or catching-up with Bertrand? I have an interest in both hemispheres of computing - academic and commercial. And in this case, Bertrand wins - a few hours of discussion with Bertrand is easily on a par with any two weeks of PDC content. I'm fully recharged and off to implement Eiffel-like concurrency, contracts in Rotor-C#, proof-carrying code, and a B-to-Eiffel translator. Not to mention working on solving a great puzzle supplied by Rustan. But first, I need to pay some bills...