Blog Not Needed

In order to better get the latest news and updates out to you folks, we put a two new sections on the home page. Basically a snapshot of the blog headlines, in an attempt to draw a bit more attention to the blogs.

Why? Well because we often cover topics and issues that are of significant relevance for the site. Especially with regards to future efforts, and changes that effect users.

It is all about transparency here.

I’ve been working on some new minor features here and there and did a roll out while I waited for stuff at work. Anyway, here’s what is new. First is the Top 10 links. Basically I came up with a way to determine the “top 10″ links in the system based on a variety of variables. I’m not going to go into specifics of what, but I’ll just say that it isn’t just click thrus and impressions. No, that would be too easy to game.

Editor Picks are just that, links our editor (and me) find particularily interesting/good/relevant/important that need some highlighting. The list is a “rotating” list, meaning that as editor picks are added the list gets recycled (on the home page). If a link is designated as an “editor pick” you’ll also see that in the link details page.

Lastly, and is related, I created RSS feeds for both, and a live bookmark function (for the Top 10 links) So, point your RSS news readers here and here.

Next up is to finish DiyShare and the extended link reporting functionality.

Today’s word is: “oops.”

I was taking a little break from my actual work to push a small tiny code change into production, but didn’t realize the WAR file (the archive file that contains all of the diysearch java code and dependent libraries) had a build problem. I didn’t realize this till *after* the archive was moved into production, and after crawling through mountains of log files trying to figure out what the problem was.

Well, to make a long and dull story short, I did a re-build, and pushed the entire WAR into production.  Everything is working again.

Sorry about that.

Code rollout has been completed. Took about 10 minutes to do, but the end result is that you can now log into the blogs and the site using the same username/password. There will never be a need to make multiple “accounts” for various systems from now on.

Yeah, I really should have done this much earlier, but it didn’t seem to be a priority to me until I started getting the feedback. So, I’m going to scratch this one off the “to-do” list.

I finally managed to finish the “universal login” to integrate the blogs (this one and the editor blog) with the main diysearch user database.

While I think Wordpress (the system that drives the blogs) is a fine piece of software (with regards to features and stability) it leaves a lot to be desired in the actual implementation. This is a mess, a hodge-podge of display/HTML code mixed in with business logic. A confusing maze of some object oriented concepts intertwined with procedural code. A staggering amount of nested includes, no real clear sense of an object model and absolutely horrid documentation on the actual code itself.

Wordpress, while designed with an “API” (or a plug-in type of architecture), which, in practice, is supposed to simplify the process of “overriding” parts of the system, it is incomplete and inconsistent. Mixing variable names from one function to the next, not following good OO practices for object creation and management, using a weird authentication scheme where the username is fetched, checked if it exists, then checks the passwords? Two steps? That seems a bit strange to me.

If I had the time and interest I would replace Wordpress with a blog system of my own, but alas, I don’t and I won’t.

The code rollout will most likely happen today. I’ll  post a thing saying when its done. Basically you can sign up at the main site, and use that login at the blogs… and you can change your information at either place (on the main site or on the blog profile screen).