Comments for GrinBlog 2.0

Comment on The Costco Stamp Mystery by Christopher

Christopher — Wed, 24 Jun 2009 22:13:13 +0000

Is it possible that the US post office had just upped there postage rate and that costco was selling off their remaining stamps at a reduced value?

Comment on The Occasional Tech Comment: MySQL GROUP BY Oddities by Adam Batkin

Adam Batkin — Sat, 20 Oct 2007 23:18:00 +0000

Heh, sorry. I just never liked mysql much since it always seemed like the whole thing was a hack. And there’s PostgreSQL available for the same price (free), the same ease-of-use and administerability (easy) but it’s an actual database with real transactions and much better standards compliance.

Then I go back to work and install some piece of software and it inevitably requires mysql with no possibility of using any other database.

Comment on The Occasional Tech Comment: MySQL GROUP BY Oddities by kevin

kevin — Wed, 17 Oct 2007 15:59:21 +0000

Right… MySQL’s current behavior - where it will, in fact, pick an arbitrary value and not throw an error - struck me as slightly insane. (Though apparently my mocking tone in the summary didn’t communicate my feelings as clearly as I would have liked!)

Comment on The Occasional Tech Comment: MySQL GROUP BY Oddities by Adam Batkin

Adam Batkin — Wed, 17 Oct 2007 15:54:17 +0000

If anything in your SELECT clause isn’t in the GROUP BY clause (other than things that would already be grouped by primary key, something which plenty of RDBMS’s get wrong) then you are *guaranteed* to get incorrect results.

The database should stop you because, there is no way it can possibly answer that query correctly! The database would have to ARBITRARILY pick some random values and spit them out.

The only other possibility is that your data isn’t properly normalized and at the application level you “know” that certain values will be already grouped, but still, the database doesn’t know that (and in that case, adding those values to the GROUP BY won’t change anything)

Comment on Props to: Spoon, Amazon MP3 Store by Adam Batkin

Adam Batkin — Wed, 03 Oct 2007 09:03:20 +0000

I tried buying an album on my copy of XP running in VMware and it was quite a nice experience (except for it constantly opening Media Player even after I kept closing it). I could see buying all my music this way from now on.

Comment on Props to: Spoon, Amazon MP3 Store by Adam Batkin

Adam Batkin — Tue, 02 Oct 2007 20:41:42 +0000

Linux (and Mac) users are often the early adopters, and a very influential bunch. If you can get us hooked, whatever it is has a much better chance of being successful.

I know there aren’t any good statistics out there, but I’d wager a guess that there are probably just as many desktop linux users out there as Mac users.

Comment on Props to: Spoon, Amazon MP3 Store by kevin

kevin — Tue, 02 Oct 2007 12:17:18 +0000

Well, they’re saying that a Linux client is in the works - and since the whole thing is in beta, it’s pretty reasonable to say that we’ll get the 98% solution out first.

Comment on Props to: Spoon, Amazon MP3 Store by Adam Batkin

Adam Batkin — Tue, 02 Oct 2007 08:57:34 +0000

But they still hate Linux users. Only Mac and Windows users can buy albums from Amazon.

Comment on Well played by Marli

Marli — Fri, 21 Sep 2007 19:07:50 +0000

My entire office is loving this article.

I can’t believe this is real, not The Onion.

Comment on From the Department of “Duh” by Yoni Samlan

Yoni Samlan — Tue, 04 Sep 2007 19:10:39 +0000

the sample size is only 1000, though, with only 100 deaths, and their geographic studies between american and british artists are on even smaller samples. Throw in the scattershot timeline… well, i’m sure there’s a pretty sizeable margin of error on this one. Interesting anyways. And I definitely would be curious if the rate of deaths decreases heavily for certain types of bands - for example, members of vegan straightedge bands like AFI will probably live for a very long time given the major causes of death cited were drugs, alcohol, and heart attacks. And correlations between level of fame (platinum? multiplatinum? fame peaking or at an all-time low?) at the time of death would be interesting, if not statistically significant over the teensy sample sizes.