Being the thoughts and writings of one Gustaf Erikson; father, homeowner, technologist.

Read about: books » computer stuff » mobile tech » links of interest » everything else

 

Sidebar links

I’ve relegated my del.icio.us links to the sidebar, and hidden the links category from the main page with the hide plugin. This is in keeping with other “mainstream” blog packages way of handling external links.

The solution is extremely crufty at the moment — this whole blog is aquiring behind-the-scenes cruft at an alarming rate. I hope to do something about it soon, in my copious free time.

Automatic linkblogs

When it comes to maintaining this blog, I’m trying to stick to the DRY principle expounded in The Pragmatic Programmer (“don’t repeat yourself”). I’ve recently discovered del.icio.us, and it’s every bit as cool and tagalicious as everyone says.

It’s pointless to grab URLs and post them here as linkblog entries and then do the same thing on del.icio.us. That service is way better than anything I could come up with, and it has an API. Keep it simple, and don’t repeat yourself.

To that end, I’ve spent a happy evening coding a little perl app that’ll grab the latest entries from my del.icio.us account. If they’re tagged with linkblog, they get posted here too. One place to enter the data. Maintaining it will be a bit different, there are some things, such as attributions, that can’t easily be handled in the del.icio.us interface. But I can always fix that later.

The app is not ready for prime time yet. I’ve based it more or less blatantly on this code but tweaked it for my setup. I’ll post a link when it’s finished.

Throwing in the towel (nearly)

Michael nearly pulls the plug on his blog. Low-life spammers are responsible, of course. This actually makes me feel less bad about pulling comments from this blog. But it still sucks.

Clearly, something has to be done. I see a great need for a Bayesian filtering system for blog comments. This has worked wonders for my email spam, and is a really good weapon to have in your arsenal. But I haven’t seen anything like it yet, at least for Blosxom.

(Michael’s system, Movable Type, has its own problems with serving pages. But the base problem is the same.)

I’m thinking about writing my own comment submission form that’ll use SQLite or Berkeley DB to process the raw entries, apply Bayesian statistics to them, and present them nicely for moderation. Blosxom’s file-based layout has obvious drawbacks when it comes to rapidly handling lots of data from different angles.

When I’ll have time for that is another matter.

Comments are off again

I’m fed up with dithering with comments and trackbacks. I’ve removed everything for the time being. If you have a pressing need to tell me something, see the contact page for directions.

I could insert a rant about how commenting should be an integral part of blogging and what a pity it is that spam is poisoning the commons, but the fact is I have no visitors and nothing interesting to say anyway so why should I bother with comments?

I’ll try to get a real hang of how the new-and-improved commenting plugins work with blosxom in my copious free time.

First year anniversary

Today it’s been one year since I started blogging. My first post was a “review” of the Lord of the Rings. Since then I’ve written 49 capsule book reviews (one for each book I’ve read), used three different weblog packages (one homegrown, Movable Type, and now Blosxom) and basically bored more and more people each day.

It’s been fun until now, and I’m definitely keeping it up for a while. Resolutions include:

  • getting comments working while defending the blog against spam

  • learning more about Blosxom and helping out on the mailing list

  • maybe writing my own plugin (don’t know what it should do yet, though)

I’ll also try to write better about interesting stuff and less about blogging.

New category: scrivener

Updated.

I found that I had a lot of site-specific posts about the ongoing re-design of this here blog, so I thought I’d experiment in creating a new category.

This is how I moved my entries:

  1. Got the redirect plugin from Fletcher Penny.

  2. Found the entries that were site specific from the original category, and listed them in a file for future reference.

  3. Created a new directory for the category.

  4. Moved the files to the new directory.

  5. Added the moved files to the redirect configuration file (see below).

Thanks to Doug Nerad for pointing me to the redirect plugin.

Anyway, now all site-related entries will be found in the scrivener category. Enjoy or avoid, it’s your choice.

Ugly XML icon

Some people want us all to use the ubiquitous XML icon. Other people think it’s ugly. And some of us wouldn’t care one way or another — myself included — if we didn’t know that Dave Winer is responsible for it.

Anyway, I’m replacing that icon with a blogbutton: RSS 1.0
blogbutton. But before I banish it into the bit bucket, I’ll share with you how it’s made. It’s not an image, but a bit of CSS:

<a title="RSS 1.0" href="http://gustaf.symbiandiaries.com/weblog/weblog/index.rdf"
   style="border:1px solid; border-color:#FC9 #630 #330 #F96;
   padding:0 3px;font:bold 10px verdana,sans-serif;
   color:#FFF;background:#F60;text-decoration:none;margin:0;">XML</a>

that looks like this when rendered: XML.

Cool huh? I got it off the net somewhere along the way. Thought I’d share. If someone owns up to claim authorship, drop my a line.

I might add that my objections to the icon is not primarily aesthetic (after all, the replacement is orange too, but that’s OK, because orange is the new black) but functional. The XML icon implies something else than a link to a syndication feed. I must admit I was puzzled the first time I encountered it. Then I found out what it was for. But it is puzzling. The new icon at least advertises what it is in a better way.

Commenting redux

Updated

Well, as with many technical problems, the solutions present themselves after a night’s sleep.

There was no problem with the URL rewriting (I had simply disabled the writeback plugin previously, and neglected to enable it last night while testing). These are the steps I followed to enable trackbacks.

  1. Install and configure the Blosxom Writeback plugin. There are clear instructions in the package.

  2. Download the stand-alone Trackback implementation from Movable Type.

  3. Copy the tb.cgi, header.txt and footer.txt to your cgi-bin directory.

  4. Edit the tb.cgi file. I simply replaced the $Password variable with a good password.

  5. Create the directories tb_data and tb_rss in the same directory. (To be wholly honest, I don’t need if these directories are needed.)

  6. Enable the display of the trackback URI where you want it (check out the flavours included in the writeback package).

All done!

I’ve tested this a bit, and as far as I can tell, I don’t get a mail from wbnotify when a trackback is received. So something will be needed to keep track of what’s posted.

Original post

I was forced to disable comments a while back because spammers were making this blog into a cesspit — as they’re making blogs all over the planet as we speak.

I’ve been working with enabling comments or even better, enabling trackbacks. Trackbacks are not immune to spam, but I agree with oblios in that a weblog is a publication. By at least requiring the commenter to have access to a blog of their own, you raise the bar slightly.

All well and good, but unfortunately it seems that blosxom and trackbacks are compatible, not many people seem to have implemented them. In particular, I found very little info on the nitty-gritty of how to configure blosxom’s writeback plugin and the stand-alone trackback CGI. I found an additional wrinkle, too. I use Apache’s mod_rewrite to translate /cgi-bin/blosxom.cgi to /weblog/, and I think I need to hack writeback to reflect this.

All in all, a slightly frustrating experience. I love blosxom and like its philosophy, but the technological laizzes-faire model of decentralized plugin development can feel sub-optimal at times.

Linkblog

I’ve started posting links with or without short commentaries. They’re the ones with arrows instead of titles.

You can, if you’re so inclined, subscribe to just the links by using this link:

http://gustaf.symbiandiaries.com/weblog/links/index.rdf.

However, there’s no way of subscribing to the main feed and not the links. Deal with it, or write a filter ;-).

XHTML-MP

Gotta get this blog mobile-friendly, maybe XHTML-MP is the way to go.

Got these references from Tarek:

  • O’Reilly tutorial, pt 1
  • O’Reilly tutorial, pt 2
  • Openwave tutorial
  • Introducing WALL

Update: I added the XHTML-MP DTD to the page, and everything seems to validate. I’ll check it out on the mobiles I have later.

Unicode characters

I’m a bit torn about how to handle the “meta-strip” below each post, the one containing the posting date, permalink, and so on. The octothorpe (#) is almost universal for denoting permalinks. Some people have recommended the ‘paragraph sign’ or pilcrow (¶) instead, but I’m not happy with that in Verdana. I’m going to try with the ‘N-ary product operator’ (∏, ampersand notation: ). The big Pi suggests P as in permalink, and also the kind of Grecian edifice that stands the tests of time.

The small pi is included here for possible future use: π (π).

I got the ampersand codes for above from this page, which weirdly is a subset of a Jane Austen-oriented site.

The vertical bars separating the fields was getting too bold, so I’ve replaced them with non-breaking spaces. I’m looking for a good, unobstrusive character to separate the fields. Maybe I’ll just style the bars differently.

Update Digging around on Alan Wood’s Unicode resource site, I found the following interesting candidate for permalink characters:

  • hash/octothorpe: # #
  • pilcrow: ¶ ()
  • n-ary product: ∏ ()
  • small letter pi: π π (π)
  • lozenge: ◊ ()
  • nabla: ∇ ()
  • reference mark: ※ ()
  • double-struck capital P: ℙ ()
  • strictly equivalent to: ≣ ()
  • place of interest sign: ⌘ ()
  • OCR belt buckle: ⑄ ()
  • OCR fork: ⑂ ()

Under construction

I’m gonna do a pretty radical redesign (or undesign) of this blog.

As no-one reads this anyway I figured I could do the changes on the production sytem, rather than mess around with testing and staging.

So if it looks weird to you, this is the reason.

Update: done, for now. Some tweaking to do, but overall, I’m pretty happy.

The main points fixed are:

  • moved the sidebar to the foot of the HTML source. This means that the content will be shown before the nav stuff when using w3m, links, or lynx. The techniques in the article “Creating Liquid Layouts with Negative Margins” were used to accomplish this.

  • I’ve improved the semantics of the layout, with real h3 headings instead of just strong tags.

Site update

I’ve updated the design a bit (see the footer for attribution). I like it so far, but may try to make some changes in the next couple of days. At least I’m not ashamed of the layout anymore.

I’ve implemented the Blosxom meta and interpolate_fancy plugins (see the colophon for details). This means that you can now see when a post was updated.

Fiddling with CSS

I’m dinking around with a new CSS stylesheet. The changes will mostly be internal. I got my inspiration from Frank Hecker’s well-designed stylesheets, along with the pointer to the liquid 2-column layout detailed in A List Apart.

The first design for this blog was a OSWD design called Oggle. I found OSWD to be a great resource to get going with CSS, as you get something that you can start with and hack around until you’re happy. But ultimately the cut-and-paste without knowledge of what I was really doing started bothering me.

The design as it is now is usable, but not something I’m really happy with. I’ll try to fix things in the coming days.

(By the way, I think this is my 300th weblog post. Go me!)

Double century

Wow, somewhere along the line I passed 200 posts. This will be the 207th.

There’s the drivel-generating power of emacs for ya.

Dealing with comment spam

Let’s face it: it’s a war we can’t win. But in the meantime, here’s how I handle the (modest, for now) amounts of comment spam on my site.

I’ve set up wbnotify to mail me when I get a comment. When spam arrives, it’s usually consistent in the form of included URLs, i.e. the same link is posted more or less at the same time.

I got a script called blog-grep.pl from somewhere (if someone recognises this as their handiwork, please contact me so that I can attribute this correctly). This script makes it very easy to search your writeback files for the offending string, and to optionally delete them.

This solution is dependent on you having command-line access to the writebacks themselves, but I suppose it can be used “offline” if you download the files via FTP and run the script locally.

Disabled comments

My sanguine views about dealing with comment spam have proved to be too rosy. I’m hit bad by idiots posting spam. So I’ve taken comments offline until I can find an effective way of dealing with this shit.

Summary of the state of play so far.

The worst form of blogging

… is the pointless day-to-day diary of your daily doings.

If you can read Swedish, you can read my form of this sin at huset, my daily recap of my “vacation” working on our house. (I’ll leave the fact that it is impossible to afford to pay a professional to do stuff on your house in Sweden for another rant.)

My defence of this practice is that I want to try it out, and also that random thoughts occur to me when I’m sanding a wall or whatever, and I think: “I’m so blogging that”. (Of course, by the time I turn on the computer in the evening I’ve forgotten all about it.) This helps me through the drudgery of manual labour.

Also, I rather like the idea of a free-form database of info like what colours we’ve used on the walls.

But I’m painfully aware of the blog-wankery involved … we’ll see if I’ll keep it up.

Pictures at Mr. X

MrX Photographers is a site devoted to digital photography. Terje, the guy behind the site, is a Mobitopian and all around nice guy.