Google Reader’s Nomenclature

I really like Google Reader as an RSS aggregator and use it regularly. But a couple things still annoy me.

One is that there is no way to set the default feed order (oldest first or newest first). But whatever. That only has to change once per feed. What really puzzles me is how they refer to organizing your feeds.

On the “Manage Subscriptions” page, you’re allowed to create and add feeds to “folders.” From the main page, you can see these folders (with little folder icons) with feeds underneath them.

In the settings area, there is a tab for “Tags” in which you can delete and manage your tags, which are the same set as the folders from the previous page. Each feed post has a list of tags at the bottom. And the search boxes use the word tags.

To add to the confusion, they used to use the word “labels” for these same tags. In fact, the URLs for tags use the word “label” and pressing g+l brings up the tags list. (And gmail uses the word label for the same concept.)

Not a big deal, but it seems odd to consistently conflate the idea of a flat tagging system with a hierarchical folder layout. I’m almost positive that not long ago, they had all three names coexisting.

Google Finally Groks XML

a google search for my name

For a long while now, I’ve bemoaned Google’s lack of proper XML-parsing.

It’s done all sorts of half-way steps, gradually going from showing blank text and no cache with an “unknown” media type to caches to plaintext versions with bare tags.

It appears that Google finally understands my application/xml pages. I don’t know if it goes all the way to supporting application/xhtml+xml, but I’ll take what I can get.

This means that if you use the recommended method of serving XHTML to all half-modern browsers without any goofy browser sniffing or text/html, you won’t be left in the web dust. Yay!

Footnotes for WordPress

an asterisk

My penchant for footnotes ((Like this one)) has been recorded, so it’s no surprise that I often wish I used them more.

Online footnotes are somewhat obviated by links ((But not totally — asides are a lost art)). Now, they are easy enough to do stupidly ((Just put a little asterisk in and a line at the bottom of the post)), but if you want to create linked footnotes that go to the footnote and back, things get a little tricker ((For one, the links have to be unique across one’s entire blog, since any number of posts can be on the same page)).

So I searched around online for a good footnote plugin and found a decent one. Unfortunately, it used numbers, not symbols. I like symbols ((As was noted in the aforementioned post)).

Anyway, point is, I hacked up his code to add symbol support and am making it generally available ((The original license was very vague — seemingly public domain, so mine is too)). I tried to make the addition general purpose ((And I’ve included a symbol override option, if you don’t like the awesome defaults)), but it’s somewhat a hack.

One thing that I was a little surprised ((Alright, not that surprised)) to see missing from CSS was the ability to set custom strings for list items (for my symbols rather than numbers). But oh well, I just hide the list item markers and insert the symbol. Again, it’s gimpy, but you may gain some use of it.

RDFa

I’m so pumped right now! I just stumbled across the RDFa Primer 1.0, and I’m aglow about the future.

I’ve long been interested in the Semantic Web, RDF, FOAF, and such. The whole time, I’ve pined for a way to more closely marry XHTML and RDF. Doing so would make it more feasible to author in the first place and easier to maintain. I have a FOAF file, but it still lists my job from a year ago.

The primer describes RDFa, which is basically a collection of best practices and examples for how to describe RDF triples inside the constructs of XHTML 2.0. This would let us do neat stuff like:

<p class="contactinfo" about="http://example.org/staff/jo"
       xmlns:contact="http://www.w3.org/2001/vcard-rdf/3.0#">
        My name is
        <meta property="contact:fn">
            Jo Smith
        </meta>.
        I'm a
        <meta property="contact:title">
            distinguished web engineer
        </meta>
        at
        <a rel="contact:org" href="http://example.org">
            Example.org
        </a>.
        You can contact me 
        <a rel="contact:email" href="mailto:jo@example.org">
            via email
        </a>.
    </p>

Unfortunately, it requires XHTML 2.0. Or maybe fortunately. I always like playing with new stuff, and I’m getting sicker and sicker of supporting IE. Hurry up, W3C! Finish the specs.

Fickle IE

I’ve been using for some time now a workaround to serve my web site as XML to all browsers, even IE. This gave me warm fuzzies because I was serving super cool XML without any gross browser sniffing.

But recently, it seems this trick stopped working?! I only get a blank page when attempting to access mterry.name from IE. I’ve made simple test cases, and it definitely is related to just that trick. Served as XML, IE will display the DOM tree. Once I add the XSL copy tranform, it gives a blank page.

It doesn’t seem to only be me. Dean Edwards has a sample page using the trick. That also no longer works.

And it isn’t a case of IE being updated, I don’t believe. This happens even on my roommate’s ancient Windows 98 machine with IE 5.0, which he religiously fails to update.

What’s the deal? Can anyone explain this?

I’m tempted just to move completely to application/xhtml+xml and forget about IE. It’s such a decaying waste of time, and the one person that complained only had to use IE at work. It seems my target audience (friends and family) all use Firefox or some such. Good for you guys.

Update: I fixed it. A <script> tag in my <head> section was self closing. Making it like <script></script> let IE deal with it.

And IE 5.0 apparently never worked anyway. Sigh. One of these days, IE! Bang, zoom, straight to the moon!

HTTPS

You may or may not know about HTTPS, the secure, encrypted version of the normal protocol used for the web. Rather than transmitting your passwords and surfing trail around in plaintext, it’s all encrypted.

Normally, sites use it for logging in, to protect your username and password. This is great. But why don’t we use it all the time? There is some cost on both ends to encrypt the data, but surely the cost isn’t prohibitive?

For most sites, it doesn’t matter, but a recent egregious example is Gmail. They’ll use HTTPS to log you in, sure, but after that, they downgrade you to the same protocol on which you came. Which means that all your email is passed around in plaintext if you didn’t explicitly ask for encryption. You can turn on encryption by just changing the http:// to a https:// in the location bar. There’s even a greasemonkey script to force Gmail to use HTTPS. But why should you have to bother?

Another silly example is FastMail, my erstwhile mail provider. On their front page, they have two login buttons: “Secure Login” and “Login.” These buttons swap location depending on whether you access it from a https:// or a http:// address. Why would anyone want a non-secure login?

I think these sites should just force encryption. Especially since they are both email services which traffic in very sensitive information. Other sites could use it too, though. I don’t see why I wouldn’t want my shopping, searching, porn watching, or really anything at all encrypted.

For my part, I will try to use the https:// form of addresses in my links when I can, starting with this post.

Google Slightly Better

Google has historically had terrible support for my websites, which are served as application/xml. My pages would show up without a summary, an unrecognized file type, no cached version, and clicking on View as HTML wouldn’t work.

Now, it seems Google is working on XML support a tad. Some of my pages show a summary and a working View as HTML option. No official cache (though, the HTML-ized version is a cached copy) or file type recognition yet, which means they probably still don’t parse or follow my links.

www

OK, folks. The www prefix was cute in 1996, but get real: there is no other web besides the world wide one. We all know that gooberface.com is an Internet address without it.

If you really need to be precise and say, Access this with a web browser, then just include the designed-for-that-purpose http:// prefix.

www is completely redundant and therefore imparts no information whatsoever. The world could be saved some bandwidth, typing time, and ease-of-pronouncing-addresses if we all just stopped using the prefix. Thank you.