This Saturday, I'll be at Powell's Books with a bunch of other Perl authors (for those who don't know, my real name is Curtis Poe). I find this amusing for a few reasons. First, the other authors who will be there are better known and respected than I. Second, my book isn't even out yet (and I'm only third on the cover -- I contributed less than the others). Third, and this is the biggie: no one told me I was going to be there.
Last night I received an email from one of the coordinators. He was trying to plan things, wanted photos for flyers, etc. That's when I found out I was on the list. Surprise! I don't mind since I'll be free on Saturday, but I admit that I'm caught a bit off guard.
And since I'm a geek, I must confess that I noticed something interesting about the Powell's Web site. It's probably a bad interesting and not a good interesting, but then, I don't know the technical reasons for the decision.
Let's a take a look at one of their hyperlinks:
<a href="http://www.powells.com/biblio/059 6100922">Perl Testing: A Developer's Notebook</a>
Hmm, what's that number in the URL? The link goes to this page and that number is featured prominently. It's the ISBN of the book. The use of that number immediately struck me as a potentially bad decision because you don't want to use identifying information for links like this. When new editions of books are put out, the ISBN number sometimes change (and they definitely change when a different publisher puts the book out).
_sister_madly_ points out that The Great Gatsby has been reissued so many times that it has many, many ISBNs.
I suppose they did this because you can easily link to a specific edition of a book. This is handy if there's a new edition you want to highlight. However, what happens if you have 20 ISBNs for a book, you have books in stock for 19 of those ISBNs but not for the ISBN you linked to? What if the book was published before ISBNs were introduced? (1966) Worse, what if an ISBN needs to be changed (perhaps it was entered incorrectly or the book is recalled and reissued)? Then the ISBNs on the Web site might need to be changed in multiple places rather than a single place in the database. If the hyperlink had simply used a non-identifying ID (a UUID or a database ID with an optional ISBN for a "default" edition), these problems would simply go away.
Of course, it's quite possible they've considered and eliminated all of these problems, but I strongly suspect they didn't (because I see mistakes like this all the time). If they did anticipate these problems, they probably had more work to do since they use identifying information in the first place.
Last night I received an email from one of the coordinators. He was trying to plan things, wanted photos for flyers, etc. That's when I found out I was on the list. Surprise! I don't mind since I'll be free on Saturday, but I admit that I'm caught a bit off guard.
And since I'm a geek, I must confess that I noticed something interesting about the Powell's Web site. It's probably a bad interesting and not a good interesting, but then, I don't know the technical reasons for the decision.
Let's a take a look at one of their hyperlinks:
<a href="http://www.powells.com/biblio/059
Hmm, what's that number in the URL? The link goes to this page and that number is featured prominently. It's the ISBN of the book. The use of that number immediately struck me as a potentially bad decision because you don't want to use identifying information for links like this. When new editions of books are put out, the ISBN number sometimes change (and they definitely change when a different publisher puts the book out).
I suppose they did this because you can easily link to a specific edition of a book. This is handy if there's a new edition you want to highlight. However, what happens if you have 20 ISBNs for a book, you have books in stock for 19 of those ISBNs but not for the ISBN you linked to? What if the book was published before ISBNs were introduced? (1966) Worse, what if an ISBN needs to be changed (perhaps it was entered incorrectly or the book is recalled and reissued)? Then the ISBNs on the Web site might need to be changed in multiple places rather than a single place in the database. If the hyperlink had simply used a non-identifying ID (a UUID or a database ID with an optional ISBN for a "default" edition), these problems would simply go away.
Of course, it's quite possible they've considered and eliminated all of these problems, but I strongly suspect they didn't (because I see mistakes like this all the time). If they did anticipate these problems, they probably had more work to do since they use identifying information in the first place.
- Mood:
thoughtful

Comments
Sounds like fun - wish I was going to be in the area. Good luck with the event. If your book isn't out yet, what are you going to be doing? Are you guys talking on anything, answering questions, whatever (signing 8X10 glossies *laugh*)
Yeah, has it really been 5 years? Had to be, because I didn't have kids then.
If you get to DFW for more than a couple of hours, or if I ever happen to be in an area you are again, we'll toss a few back. Maybe do some dancing. Wheee.
Hope all is well with you!
Programming does not bore me, although I try to be blissfully ignorant. I am sure that would be no obstacle to conversation with you though. Perhaps some other time. Keep posting info on what you are doing and I may just show up one of these days.
-B