random bits

[cellio]

Having completed the first pass at digitizing or replacing our folk music on old media (we still need to do some proof-listening), Dani and I are merging our iTunes libraries so this might be easier going forward. Oof. We're up to "S" so far. "T" is big because it includes all the "The"s. Tracking changes (e.g. to tagging) going forward is still going to be a bit of a challenge.

Was Joe Biden president of the US for about 5 minutes today? (We were watching in a conference room at work, and it was several minutes past noon before they got to Obama's swearing-in. So I'm curious.)

In English we say "it's all Greek to me". What do speakers of other languages say? Whom do they implicate? Wonder no more; Language Log has a nice graph of some of these. I admit to being surprised by China's designee.

What if the stop sign were designed by corporations? (link from filkerdave)

As dsrtao said, an airline charging a cancellation fee when they rebooked you on a downed flight is near-canonical chutzpah. (Yes, I saw the note that they recanted.)

This story of a mailing list gone wrong (from Microsoft) made me laugh. And sigh, because while I haven't had to deal with quite that level of mess, even 20ish years after mailing lists started to become broadly accessible, there are still an awful lot of people out there who don't behave appropriately.

There's an interesting discussion of filtering and politeness on social networks over on CommYou.

Note to self: if Shalom Hartman Institute is too expensive this summer, the Aleph kallah might be an alternative. It could be good or it could be too esoteric for me; I can't tell from the available information. When they post class descriptions I'll have a better idea. I had a similar concern about NHC but it turned out to be good, so I'm keeping an open mind. Has anyone reading this gone to one of these?

Comments

[goldsquare.livejournal.com]

Email - it's a stupid software design, and nothing more.

I say that because: we know how people behave, and then if one builds a system that absolutely frags itself when people do what people do, the system is ill-designed.

Not that Microsoft hasn't made a lot of money attempting to retrain the human race.

[geekosaur]

Honestly, most of TCP/IP is that way; the protocols were designed for a tiny ARPANET.

[goldsquare.livejournal.com]

Please explain? I do not find it so.

Perhaps that's because they keep generating new RFCs at IETF. (And, for the nonce, let's ignore IPv6...)

[geekosaur]

Telnet? FTP? (for that matter, TFTP?) rsh/rlogin? DNS's various weaknesses, which DNSSEC doesn't fully address in part because it would break compatibility? There have been a lot of bandaids applied, but it originated in a small, controlled environment and that still shows in key places.

[goldsquare.livejournal.com]

I think I see your point, but if you would forgive me, I think your example is a poor one. TCP/IP worked, and continues to work - and is not against type or against the way people behave. If it weren't for address space shortages, it would remain fine.

Some of the protocols layered on top of it, made assumptions about the use of the Internet that were naive - mostly the assumption that trust can be levied. (Hell, I remember, barely, when Stanford would let you create an account on login...) In some respects, the failures you are citing were not so much retraining the user in how to use the system, but in exploiting people's natural abilities to trust.

The examples could be trust-hacked, and some simply did not scale well, but they were not always examples of "do things in an unnatural way, or your system will hemorrhage and die".



As for your examples, just to fellow-geek...


Who uses those? SSH and SSH2, SFTP. And DNS is fast evolving + there are many alternative implementations including djbdns. Many of the commercial DNS providers use proprietary DNS. Until a recent layoff, I was at Akamai - and the customer facing portions of Akamai use a home-brew DNS that has so far been immune to all published DNS flaws I am aware of. (Some systems did use bind, but it would have done little good to anyone to pervert those machines - presuming they could.)

The universe of IETF is full of examples of tried and failed/replaced or tried and adapted protocols. But (perhaps with the exception of MIME, maybe others I am not thinking of) there is little that makes users act against their own natural interests, in order to protect the network.

[dglenn.livejournal.com]

"Who uses those?"

Huh. I just posted (http://theweaselking.livejournal.com/3207885.html?thread=16088525#t16088525) a comment about the use of some of those elsewhere. I use ssh and occasionally sftp, but I still find telnet, rlogin, and ftp useful enough that I'd hate to have to do without them.

[cellio]

In my opinion mailing lists should never default to reply-all; users should have to take explicit action, and if that means a few useful messages don't make it to the list, surely it is dwarfed by all the useless messages that would have been sent by people who don't know how to behave. Alas, most mailing lists I am on are not administered in my preferred way. Fortunately, there's procmail to help.

Yes, Microsoft had some bad software design going on there (maybe still does for all I know). But people are also stupid sometimes, and wishing for a design that better defends them doesn't mean we can't also wish that people weren't stupid.

[goldsquare.livejournal.com]

You are exactly right - the software should accommodate people, as they are.

And not break when they act as they are.