Archive for February, 2006

new code

Saturday, February 25th, 2006

Work has finally slowed down enough that ive had some time this week to hack on code.

First up, I wrote mod_renice. It is incomplete and might even crash. It is designed to detect a Subversion REPORT request, that contains an update that is expensive to generate. When this condition is met, it sets the Process Priority lower, so the machine doesn’t die. The ASF Subversion repo has recently had issues with this. The thread on dev@subversion contains more info.

Second hack of the week is mod_serf. It uses the Serf HTTP Client Library to act as a reverse proxy. But, we already have a reverse proxy you ask? Yes, but the code is ugly, and this looked fun to write… so I wrote it.

Blog Software for my Mother

Saturday, February 18th, 2006

So, this isn’t one of those theoretical ‘easy enough for your mother to use’ questions.

My Mother wants to have a blog. She wants to post pictures of her quilts and stuff like that. (Yes, she is a quilter!)

So, I have looked at setting up Wordpress, or another instance of Typo for her, and that is the easy part.

What I really want is a WYSIWYG desktop editor for her. So she can insert pictures, and do all the stuff like she is used to in Windows Applications, press publish, and a whole bunch of magic happens.

I personally use Ecto on my mac, and love it. If she had a Mac, I would buy Ecto for her too. Damn, if she only had a Mac, it could integrate with iPhoto and iWeb and it would be so much easier. Maybe I will just drop-ship her a Intel mac mini, when they come out with one.

Does anyone have any recommendations on either a hosted service, or a windows desktop application that makes it damn simple to make blog posts with images?

scary

Tuesday, February 14th, 2006

It is scary when people you don’t know, take pictures of your office. If you want to come inside sometime just drop me a line. But taking pictures like some freako stalker makes me scared.

P.S.: We all use the back door to get in. Only packages come in the front. Next time, walk around to the back….

And I will strike down upon thee with great vengeance and furious anger…

Saturday, February 11th, 2006

I have plenty of hate in my heart, but I reserve a special place for websites that create freaking UGLY urls.

Today, the Ask Jeeves Blog did a post on our new Pisa Office.

They posted some photos of things going down, and hosted them on MyJeeves. i could do a whole post on MyJeeves, but lets just focus on the URLs that it generates. Here is the URL it uses to display a single image:

http://myjeeves.ask.com/share/showIm age?t=fullImage&iurl=http%3A%2F %2Fmyjeeves.ask.com%2Fservlet%2Fsni pit_image_proxy%3Fuid%3D34211032109 ec5eae5e1f25ca6e48526%26img_guid%3D 499b8b4e5c28d2de7d7c47bf62e56272%26 locale%3Den_US&user=sgrieder&#3 8;path=images+for+sharing&imgid =499b8b4e5c28d2de7d7c47bf62e56272&# 38;returl=%2Fpublic%2Fsgrieder%2Ffo lders%2Fimages%2Bfor%2Bsharing

The URL is 361 characters long. There are at least there parameters that look like some kind of UUID. Why do you need uid, img_guid, and imgid? I don’t care what the technical answer that they might have, its wrong. There are better ways to represent unique resources, and they don’t involve multiple unique hashes.

I have only been working at Ask for 6 months now. I want groups inside Ask to choose better URL designs. If you are working on a project, spend a few minutes to think about what the URLs look like. Its really not that hard.

Now, having a URL with 361 characters is pretty bad, but there is one thing far worse. That thing is MSN Spaces. Not only do they use UUIDs all over, they CHANGE THE URLs OF IMAGES OVER TIME. I have heard vaguely that it has something to do with their clustered/distributed filesystem. Its not really a good excuse to me. Sigh.

The most distributing thing is that these aren’t little podunk websites. These are major corporations, doing major projects, and making some of the worst URL designs ever.

You want to do the right thing? Go Read Cool URIs don’t change. Then read User-Centered URL Design. Good. Now don’t write code that creates those ugly URLs ever again.

Remember: People don’t make ugly URLs, bad programers do.

EDIT: Try to wrap the 361 character URL so it doesn’t scroll the entire page…. Arrrg.

public information.

Wednesday, February 8th, 2006

Bloglines keeps a running count of the number of articles it has indexed on its homepage. (Under the Search Icon, in the middle of the page… Yes, its not always an obvious place.)

1,000,000,000 Articles in Bloglines (Jan 7)

1,119,448,188 Articles in Bloglines (Feb 7, Today.)

That means in the last 31 days, Bloglines has indexed 119,448,188 articles.

This is an average of 3,853,167 articles per day.

Technorati has done another one of their State of the Blogsphere posts.

They say:

Technorati tracks about 1.2 million new blog posts each day, about 50,000 per hour.

1.2 million vs 3.8 million. This means Bloglines is indexing at least 3.2 times as many articles, every day, assuming Technorati is quoting a true Average, and not a mode or weekly peak. This puts Bloglines at about 160,000 posts per-hour.

I would also argue that Bloglines is indexing better data, since people don’t want to read Splogs, so they don’t subscribe to them, so we don’t even crawl them.

Technorati is right about one thing though, that the ‘Blogsphere’ (what a horrid word) is growing. Growing very quickly.

I am just not very impressed by their numbers that they are spouting on their Blog.