afongen

When you want peanut butter toast and you’re all out of crunchy peanut butter, but you do have that very, very large container of creamy PB, which is quite disappointing because you really don’t like creamy, but hey you’ve got that bag of walnuts so you chop up a half-handful and sprinkle it on the creamy peanut butter you decided to put on your toast anyway, cuz hey why not, don’t be all that surprised when it isn’t as satisfying as the real thing.

Still, better than creamy peanut butter. Ugh.

I’ve been working with a student intern lately, sort of showing him the ropes while at the same time picking his brain about what’s covered in his curriculum. It’s interesting (and somewhat disheartening) to see the disconnect between what I consider essential and what’s actually being taught; although to be fair, I think that his is not strictly speaking a web development curriculum. More on this later.

At one point early on, we had a long chat in which I tried to convey my strong conviction that any web developer or designer worth their salt should be more than passingly familiar with basic web standards like XHTML and CSS, and of course accessibility.

Considering how often I’ve been having this conversation lately, I started putting together a list of resources I consider essential, to serve as both introductory material and a handy reference. Working on the list has been a valuable experience: it’s really driven home how important it is to have a solid foundation in these basic, core competencies, and how often I mistakenly assume that everyone shares this common base. (It’s also helped me clean out my bookmarks.)

No sooner had I begun work than Mark Newhouse started writing about these very things in his blog, then combined the separate posts into a tidy little article on Real World Style: A Three-Legged Stool.

So there it is. Couldn’t have said it better myself.

Well, almost. Neil Gaiman writes that a “young lady named Anneli in Sweden” wrote him a fan letter:

The address she wrote on the front of the envelope was “The author Neil Gaiman. Lives in a big house of uncertain location in Minnesota USA”. On the 20th of November the United States Postal Service delivered that letter to me, care of DreamHaven Books, 912 W Lake St, Minneapolis MN 55408.

The bit in italics is from one of Gaiman’s books. Dreamhaven is a local science fiction/fantasy/comic book store that deals quite a bit with Mr. Gaiman, including creating a rather ugly online store dedicated to his books ‘n’ such. For someone at the post office to have made this connection is quite remarkable.

Either that, or Neil Gaiman is indeed Santa Claus.

A sales rep at Books24x7.com convinced someone at work to do a free trial of their online book service, and somehow I ended up in the test group. Since I already subscribe to Safari and am familiar with this type of service, I figured I’d have something to say. And sure enough… My informal evaluation follows:

Books24x7 is a good idea, something that I would probably use were it not for the fact that it doesn’t include the books that I consider the standard references in the technologies that I work with now or am likely to work with soon. Looking through those areas, I’m generally disheartened by the selection. As a quick glace at my bookshelf will attest, most of the definitive books in the areas in which I work are published by O’Reilly. They are completely absent from Books24x7, as are some other important publishers.

New Riders is thankfully included in both services.

Before you settle on this service, I suggest that you at least take a look at a competing online book service: Safari. Created by O’Reilly, it obviously includes their books, but also offers a wide selection from publishers — including many also offered through Books24x7.

I’ve been subscribing to Safari for a while now and am happy with it. I use it for reference while working, to review a book to decide whether to buy a print edition, and to learn new things — often in a hurry.

Safari’s service is basically the same as Books24x7: complete text and images, bookmarks, annotations, full-text search, etc., and have a similar interface. Two major differences:

1. In Safari there is a limit to the number of books you can access at any one time. For instance, I subscribe at the 5-book level. When I put a book in my bookshelf, it has to stay there for at least a month. Once I’ve reached the maximum number of books in my collection, that’s it until the month is over or I buy a higher-level subscription (which I can do on a temporary basis). From what I can tell, Books24x7 doesn’t have this limit, which is pretty nice. On the other hand, so far the limit has not been a problem for me: I subscribe at the 5-book level because that’s all I ever need.

   Then again, I spend a hell of a lot of money on books.

2. The quality of books offered by Safari is much, much better.

Whether Safari is a better fit for ITS, I don’t know. It’s at least worth looking at before making a decision. Would I use Books24x7? No: I already have something that’s a better fit for me.

On the other hand, I will use it to read Microsoft ASP.NET Step by Step. Because I confess to having seen some things that make me curious.

Hm, this could be a problem. The US Patent and Trademark Office granted AOL a patent in instant messaging.

So maybe AOL will go after MSN and Yahoo! for licensing fees. Bad news, lousy for interoperability (which, granted, has always been far from AOL’s goals), and especially bad for Jabber, who have a site set up to track the issue. (via ex machina)

Update: possible prior art?

Or at least his crew does. It’s amazing to read just how integrated Perl has become throughout the whole Santa organization. I always figured Perl would make sense for Santa, who’s got to have some serious need for flexibility, power, and speed in his IT infrastructure — not to mention fun — but I guess I didn’t give enough credit to his staff and assumed they’d gone the Java route. Good for them!

Hah!. And hah!

Marc Hedlund writes about Why JSP Sucks So Hard. In a nutshell, separating Java code from HTML is on the right track but incomplete: you’re still left with a bunch of stuff that makes sense neither to HTML jockies or their standard tools.

I’ve been thinking this with my early forays into Smarty. I certainly prefer working with Smarty to mixing up all my PHP code with the markup, but it’s still way too easy to build business logic into the template, something that I’m trying to avoid by using a template system in the first place. Not just on general principle, but because of the confusion faced by the designer or whoever’s now dealing with this little mini-language.

Microsoft’s made some steps forward with ASP.NET, in which the template is a valid XML file (reminiscent of taglibs, if my incomplete understanding of both technologies serves me right), but the non-HTML <asp:...> tags in the are still going to throw off standard tools. Conveniently, Microsoft’s got plenty of tools to use.

Zope’s Template Attribute Language uses attributes instead of tags, but it is still a language, embedding logic into the template. Maybe I’m being too purist about that, but that really is the point of the rant, so bear with me.

Marc suggests something akin to what was done with CSS:

Give me a “code” attribute that I can add to any HTML tag, and overload the “id” attribute. Have the value of the code attribute be an opaque string referring to a codesheet, linked to the HTML document by a a ‘<link type=”text/ccs” rel=”codesheet” href=”code.ccs”>’ tag. Have the codesheet use a CSS-style rules specification for indicating which Java (or whatever) class should consume tags with that code or id attribute. Then tell the HTMLer to add the “code” and “id” attributes to the right place in their existing documents, and they’re done.

I think he’s on to something. We’d need to work around some namespace issues (the code attribute has a specific purpose and place; this ain’t it.), but things like TAL have already pointed the way. What I really like about his idea is that, as he points out, it doesn’t bind the HTML language to any particular programming language. This would be a huge boon to development, especially if you use something like Perl or PHP for quick prototyping but have to use Java in production to assuage someone’s concerns about stability or speed. :-) Just switch out the code in the back end.

Persistent database connections are a great idea, but in my experience often used where they don’t belong. The result: reduced performance. I can understand the appeal of a perceived free performance enhancement, but there are trade-offs to keep in mind.

MySQL AB has published a little article about using persistent connections with MySQL, discussing some of the things to consider before implementing persistent connections. MySQL’s connection handling is often fast enough that they’re not necessary.

Are there any drawbacks of persistent connections? Unfortunately there are. At first each connections takes up resources. This includes MySQL internal resources (memory, table cache elements, connections) as well as OS level resources – threads, open files and sockets. Some operating systems are quite OK with this, while others may show performance degradation. MySQL resources are often more important. By having the same amount of memory that would be used for persistent connections, you usually can configure MySQL to use larger buffers, which often gives extra performance. Will this outweigh the benefit from persistent connections? It completely depends on the application.

The lesson: benchmark your application and make optimizations only when necessary and appropriate.

In case you missed it (I nearly did), Joe Clark gives a good interview for the Slashdot crowd.

He has such a straightforward, no-bullshit approach that can be grating in some circumstances but that really adds to the interview. All his writing, really. His book on web accessibility is still hands-down the best one out there. If you build web sites, read it.