Understatement of the Year - ODP "Technical Difficulties"

There was a post on SEOMOZ about DMOZ recently that resulted in some good information in the comments (I capitalized SEOmoz wrong because I like how they look together in a sentence - so sue me).

Here is the quote:

“On Oct. 20th, the editors machine crashed due to a hardware failure. During attempts by AOL Operations to restore the computer to normal operations, a disk re-image procedure resulted in the data on the editors machine to be overwritten. Attempts were made to restore the data but to no avail. Full backups to the editors machine were once automatic, but unbeknownst to us until after the hardware failure, this process was changed to a manual, on-demand process over a year ago. Only incremental backups were automatically produced.

To make matters worse, we were setting up an additional machine for failover to remove the editors machine as the single point of failure. This meant all the data on the editors machine were lost forever. In order to restore the ODP, it had to be reconstructed from various sources (i.e. tools from research, the RDF dump, etc.).

This rather extraordinary event was a bit like breaking a glass into a thousand pieces and being faced with gluing all the pieces back together. Richard (rpfuller) and David (ddrinan) assessed if the directory could be reconstructed at all by looking at pieces available in various places. There was no way for us to predict how long it would take to restore the ODP because this scenario never was supposed to be a possibility.

To add insult to fatal injury, the research server, which contained a lot of data needed to restore lost data, was taken off line due to security concerns. Richard and David were able to quickly amass the major parts of the directory. Within a week of the outage, scripts were being prepared to reload the latest RDF and incremental backups. The bulk of editor information was restored. By Nov. 8th, the editors server was back online and open to admins and technical editors for QA testing.

Between Nov. 8th and today, Nov. 29th, Richard and David have worked tirelessly to fix bugs and features reported by the testers, and to recover all data that’s possible to recover.”

It's not dead - just in intensive care...There may be some good news for it, however, as this quote from AOL to the ODP editors implies:

"Restoring the ODP to its previous state is the short-term goal, but keeping it in maintenance mode is not a long term strategy. If there is a silver lining in this outage it's a renewed interest in developing the ODP in a direction that is relevant to both the web community and AOL. Obviously, your input is pretty important to AOL in determining the ODP's future, and you guys have had conversations here and there about how you'd like to see the ODP evolve ... these conversations are not going unnoticed. As you sit here patiently waiting for the old system to come back online, perhaps you guys could begin a visioning discussion about what the ODP should become."

This is good because I think the consensus is that DMOZ as it was before it blew up really wasn't working or scalable anymore. Of course, whether or not it gets improved is up to the quality of the suggestions from the editors, and which ones (if any) get implemented by AOL.

In the meantime, I'm pretty much ignoring it. I like the idea of DMOZ/ODP, but I've been unhappy with the implementation for some time now (on an unrelated note, the same goes for Ask.com, BTW).


No comments: