August 28, 2002

More weblog metadata

It seems you can’t keep a good idea down, or something. Yet another weblog metadata project has emerged, citing the desire for interoperability among the various weblog-data-gathering resources, such as Blogdex and the possibility of new, more powerful resources in the future.

This is a concept I am fully behind, as you may have guessed if you’ve tried to read my thread description project (which suffers from a lack of a catchy name and comprehensible documentation). The care I took in designing my project (and the flexibility inherent in RDF) means that it can represent most of the metadata properties the BlogMD Initiative has proposed, and can be cleanly extended to represent the rest.

This is actually a tough situation for me: I’m convinced that the work I’ve done so far is really the simplest and most elegant way to do what the BlogMD people want (as well as to do considerably more, such as what the ThreadsML people want), but I’m reluctant to press the issue because I don’t want to come off as an arrogant prick. Also, I have yet to successfully explain it to anyone, which is frustrating because it seems so obvious to me. (via Doc Searls) #

Enter the kibibyte

While reading an article on a different subject in MWJ, I came across a reference to the proposed binary metric prefixes (previous mention). The idea behind them is that the common convention among technical types of using “kilobyte” to mean 1024 bytes rather than 1000 bytes is unnecessarily confusing (especially when you consider that hard disk manufacturers define a kilobyte as 1000 bytes). The International Electrochemical Commission proposed new unit prefixes to handle these power-of-two-based multiples. Thus, 1024 bytes would be a kibibyte and 1 048 576 bytes would be a mebibyte and so forth.

The major obstacles here are tradition and the fact that the new prefix names sound and look really stupid. The latter problem is reduced when using their symbols, such as KiB for a kibibyte, but inertia is an extremely powerful force. Creating unambiguous prefixes is a good idea, though, as right now we’re in a situation where “kilobyte” means different things depending on who measures things. Software developers might consider using KiB rather than KB when describing the size of things. It’s similar enough that it’s obvious what it means, and it’s unambiguous.

In the long term, we should reconsider the use of bytes as a unit of measurement altogether. A byte is just 8 bits, but it has no real importance other than the fact that most text files are made of 8-bit chunks. I say: why not base our measurements on the actual base? Bits, by definition, are indivisible. People in communications use bits already (kilobits per second and such), so why not simplify? After all, as it currently stands you can’t say for certain whether “1 kb” means 1000 bits, 1024 bits, 8000 bits, or 8192 bits unless you know the context it was used in. (This is sort of like the way a “barrel” means different things if you’re talking about wine, beer, or oil.)

Oh, and check out the rest of the dictionary of units. Otherwise, you’ll probably never know how many gallons are in a hogshead. #


How to make playing cards with stylesheets. (via MetaTalk) #