As his extensive suite of benchmarks and product comparisons shows, this elegantly overcomes the problem of diminishing returns that come with adding more read-only slaves in a classic master-slave RDBMS replication setup without resorting to extensive use of non-relational solutions such as memcached.
Just kidding, he did it cuz it's fun to write cutesy JavaScript hacks that leak memory like hell and turn any machine slower than a Core 2 Duo into a paperweight.
I have a piece of paper stuck to the wall behind my desk that reads: "Research is blundering into an alley to see if it's a dead end". I think this falls into the same category. It's obviously not directly useful, but it might spark a thought in the creator or someone else that does lead somewhere.
And if not, well he probably had a good time coding it.
Very cool - this is pretty similar to microformats.org's design principles. I could see this kind of HTML markup used as an alternative to maintaining a separate REST based API.
But less ambitious: it's actually not a bad way (I think - I don't know enough about JQuery to have an idea about the performance of this) to do client side filtering of data presented in tables. His technique, at least, is more readable than the kludges I've come up with to do that.
9 comments
[ 2.9 ms ] story [ 31.2 ms ] threadJust kidding, he did it cuz it's fun to write cutesy JavaScript hacks that leak memory like hell and turn any machine slower than a Core 2 Duo into a paperweight.
And if not, well he probably had a good time coding it.
But less ambitious: it's actually not a bad way (I think - I don't know enough about JQuery to have an idea about the performance of this) to do client side filtering of data presented in tables. His technique, at least, is more readable than the kludges I've come up with to do that.
In any case, this is fairly cool indeed.