I’ve had occasion over the last couple of days to upgrade our WPMU install from version 1.3 to 1.5.1 and the process was remarkably painless. It took a bit of time to get everything lined up and ready to go but the actual switch over was smooth and easy.

Largely, I followed the guidelines most helpfully shared on Robert Mao’s blog about his upgrade to v1.3. The only onerous parts were getting the DB backed up before I did anything and then getting privs to do the file-system work I needed to do.

Overall, my process was:

1. Make a backup of the whole database - there are a number of ways to do that. I used the DB copy feature in phpMyAdmin. There’s also a handy little DB backup script posted on the wordpress codex, but it didn’t work for me because the account we’re using didn’t have appropriate privileges.
2. Download the latest WPMU package
3. Unpack it at the same level as the current WPMU install. E.g. if your current install in in /web/blogs/ then unpack the latest version to /web/wordpress-new/
4. Copy form the current install into the new install:
   • wp-config.php : e.g.
     cp blogs/wp-config.php wordpress-new/wp-config.php
   • .htaccess : e.g.
     cp blogs/.htaccess wordpress-new/.htaccess
   • wp-content : e.g.
     cp -r blogs/wp-content/* wordpress-new/wp-content/
   • any custom, support directories / tools : e.g.
     cp -r blogs/latexrender wordpress-new/
5. change directory names so the new version is live:
   mv blogs blogs_orig; mv wordpress-new blogs
6. Go to your blog site and log in as the admin
7. Click on Site Admin (far right), then Upgrade (rightmost on the sub-menu), then the Upgrade Site button
8. Once the upgrades finish running check your blogs to make sure everything is running OK

I didn’t have to do any special plugin treatment or anything, but I don’t have many to worry about, and the ones I have are pretty simple. The process if problems do arise is to deactivate all of then and then re-activate them one at a time until you find the problem ones. Fixing plugin problems is left as an exercise for the reader :P

I was recently asked for my thoughts about how web 2.0 has changed the way IT support projects, specifically with the idea that web 2.0 tools required no infrastructure support on the part of the institution. I can see somewhat where that’s coming from - if the college is pushing Facebook, Flickr, Swivel, Slideshare, Blogspot, etc. then those are all products that are a) free and b) available online without the college having to do anything. In that case the support would be oriented more towards helping people understand and use those tools rather than making sure the underlying technology was working.

However, that’s an artifact of the business models behind those companies and the college’s choice to use those particular products. The shift away from infrastructure is driven more by high level decisions to outsource rather than anything inherent to web 2.0 technologies. The fact that many web 2.0 businesses provide technology support for cheap or free makes that decision to outsource more common for web 2.0 projects. In our case we decided to install and support our own wiki, blogs, and CMS. We certainly could have outsourced that (either for pay or for free), but decided against it for reasons largely related to privacy, control, and functionality.

At institutions more financially constrained I would expect to see much more of a shift. For a functionality of a certain level the options originally were:
1) build and support it in house - lots of resources, a small amount of cash
2) buy it and support it in house - some resources, some cash
3) totally outsource it - few resources, lots of cash (loss of control)

As IT skills became more wide-spread in society there arose two more ommon options:
4) use open source - lots and lots of resources, no cash at all
5) use share/free ware - some to few resources, low/no cash, but a big sacrifice in functionality

Number two is probably the most common option, or number three for projects where the loss of control is acceptable. Number five would be the preferred option if the loss in functionality were acceptable. In the past it usually has not been. The big change in the web 2.0 world is that for option five the loss in functionality is quite small.

Ten years ago the freeware version of a product might have 20% the functionality of the corporate one. Today in the web 2.0 world the free version of the product probably has at least 90% of the functionality and requires very few in house resources (for basic support). Institutions with more financial constraints (i.e. almost everyone) seem likely to decide that the 10% loss in functionality is acceptable given the large savings it provides (or rather, the ability to do more with the same amount of cash and resources). At institutions with lots of money and resources and extremely high standards that 10% loss in functionality would be unacceptable, and so such institutions wouldn’t see much in the way of change when implementing/supporting web 2.0 projects.

I tried Second Life about a year ago, and was unimpressed. However, at that time I was in a place with a not-great network connection and I really had no idea what I was doing. A bit more than a week ago Ruben Puentadura came to Williams to run a full day workshop on Second Life. I’d been looking forward to this for a while. The idea of Second Life is very appealing to me - I really want to like it. My experience with it up to that point had been disappointing, but I hoped an introduction to the world by someone already involved and interested (not to mention 12 months worth of tech improvement) would get past my initial reaction.

Unfortunately, at the end of the day that Friday I left with an even worse impression than I’d started. It’s not that there aren’t some good points to the system, but I found that the negative far outweighed the positive. Since then I’ve been trying to articulate why exactly I don’t like Second Life. In part for my own benefit, but also because most of the other people that took the workshop came away with a very positive feeling about the system and I’d like to be able to discuss at least semi-rationally why I don’t like the system and why they do. So, here’s my beef:

* The UI was terrible. The movement controls were pretty poor (offering very limited control and with very coarse * granularity), but the real problem was the modal look, move, and communicate system. I will happily admit both that this is a hard problem in general and that the IO options (mouse, keyboard, screen) are limited. But even given that the experience was bad.

* Stability was terrible. In a single day of very lightweight, introductory use of the system the client lost connectivity six times and complete-crashed twice. That’s not even getting into the lag issues and basic un-responsiveness issues. On top of these seemingly arbitrary problems the world kind of ground to a halt when a sizeable group (50+?) of people were in the same area.

* System architecture is monolithic and hegemonic - the whole virtual world exists on Linden Labs servers and there’s no way for anyone to set up their own world. This means they dictate the rules, and users are completely beholden to them. Objects in the world are useless outside the world. The corporate life of Linden Labs is a serious single point of failure for the long-term viability of the world. On the technology side, I think that also means their scalability issues are nigh insurmountable.

* The economy feels artificial. It is in many ways reminiscent of the currently struggling data-object economy in the rest of the internet, where object creation, duplication, and distribution is governed by rulings (laws in the real world, LL decisions in 2L) rather than capabilities. That is, technically speaking there’s no reason a person could not copy any object in the world, except that some objects are flagged as cannot be copied and so the 2L system won’t duplicate them. They can get away with it because they control everything, but on the flip side it only works because they control everything. It’s like a world where when a law was passed it would actually become impossible to copy certain MP3s. It’s desperately hanging on to an economic system that’s grounded in the limitations of physical production. I think the really successful virtual world will have to figure out how to avoid that artificial restriction. It’s either doomed, or it reinforces a system I wish were doomed. In either case, I don’t like it.

* The flavor / theme impositions are weird and unnecessary. For all that it’s an extremely free form world there are some very strange basic rules. All avatars must have a last name and always are referred to by their first and last name. Avatars can be killed (from which they recover, but why bother in the first place?). There’s gravity. Very strange.

Other objections have more to do with me than with the 2L itself - I react badly to a hard sell, I’m an introvert so the idea of wandering around and striking up conversations with strangers is no more appealing virtually than it is in real life, and finally, it takes a really significant commitment of time, much more, apparently, than I’m willing to give it. It’s possible that 2L will eventually evolve into the sort of virtual world system I’d like, but to me it feels more like early AOL than early WWW.

Some things that I think Second Life needs:

• much better UI - the really good stuff will probably have to wait for new peripherals, but there’s serious room for improvement even with existing technology. At a very basic level even allowing configuration of key-bindings in the client would be a huge leap forward (bringing it up to the standards of mid-90’s software).
• strong ties to the rest of the internet - built in browsers, gateways to established commerce systems, ability to handle all major media formats. In addition to making the system more generally useful and appealing, it would likely assuae much of my dislike of the economy as it stands.
• voice communication - either an independent system or tying into one of the various VOIP systems out there. Ideally letting users select their own voice support system and just providing a way for them all to hook in.
• better software - simply put, less buggy and deals with edge conditions more gracefully. It should degrade fluidly with varying connectivity, scale well to many, many users, and generally crash and lock up much less often.
• distributed system - it’s vital to the robustness and longevity (technologically, economically, and politically) of a virtual world that it not be hosted and controlled by a single entity.
• finer control - this is related to the UI issue, but a bit different. LL talks about their virtual land and items in terms of meters, but they’re not really meters in the sense that they are proportional to the real world distance. A meter might in game be represented by a distance about 1/2 as long as a tall humanoid avatar, but the effect of that distance is much different. It’s difficult to manipulate effectively anything much below a couple 10s of centimeters, and the resolution limits of both monitors and human vision end up making even small signs multiple meters long. The sorts of things humans regularly interact with - books, utensils, keyboards, steering wheels, buttons on TV remotes, cell phones, etc. - are all at a scale well below the threshold of effective use in 2L.
• permeability - need to be able to get things very easily into and out of the virtual world, whether it’s text in a virtual book, CAD created objects, images, software written in languages other than 2L script, or whatever.
• adaptability - it should be able to work with a wide variety of IO devices, and it should be easy to add more as they’re invented / available (e.g. projection glasses with momentum sensors). Open-sourcing the client software is a good step towards this.

Second Life definitely has good points as well, but I’m not really going to get into that since you can read all about it pretty much anywhere 2L is discussed and I have nothing new to add in that realm. In the end, I think Second Life is good enough for many people who want something like it, but not good enough for me to want to use it and not good enough for me to recommend it to anyone else. I don’t yet know if there are currently any viable alternatives to 2L, and if not then I’ll wait.

Here are a few other places to read some less-than-wholly-positive writing about Second Life and other software of that ilk:

http://ilamont.blogspot.com/2006/11/problems-and-promise-in-second-life.html

http://www.toothpastefordinner.com/journal/journal.php?user=toothpaste&id=573&readcomment=1

http://www.rpg.net/reviews/archive/11/11685.phtml

http://slgames.wordpress.com/2007/03/05/alternatives-to-second-life/

http://www.secondlifeherald.com/slh/2007/03/alternatives_to.html (same article, different discussion)

http://editthis.info/sl_wiki/Alternatives_to_Second_Life

Bryan Alexander came to the college today to give a pair of very good talks. If you have a chance to attend any of his presentations or classes I highly recommend it.

The first was a lunch talk on Web 2.0 and Pedagogy. It was aimed in large part at faculty, but we had a number of attendees from administrative departments as well. I think it gave people a good idea of what is Web 2.0 and why they should care, but we’ll have to wait and see what the actual response is. I expect we’ll get som faculty wanting to create blogs or start wikis for collaborative writing projects. At the very least it will serve as a good common basis for discussing blog policy on campus (most or all of the concerned parties made it to the talk).

The second was a follow-up in the early afternoon. The audience here was mostly technology people of various flavors (from a technophilic reference librarian to people in our media services group to public affairs web folks). This second talk was more interactive, with people asking questions as it went along and with the focus changing based on what people were asking about. Also, it dealt a bit more with tools and examples than theory. It was a great survey of what tools and technologies we would (or will, more likely) be dealing with as Web 2.0 sensibilities penetrate further into the academic world.

Bryan really makes me want to do and think more about web 2.0 (and game) stuff. But then there’s all this othe stuff I want to do with my time, not to mention things like my job which keeps me in paycheck. Choices, choices….

I talked with our Office of Career Counseling and our Alumni Office about setting up blogs on our new server. They were definitely interested. More importantly, it got the ball rolling on building a blogging policy. We’re still in the early discussion stage, but there are enough people with enough interest that I think we’ll get somewhere.

One of the big questions to be resolved is blog ownership / authorship. We aren’t establishing just a general blogging service here, we’re setting up Williams blogs: departments, offices, student groups, etc. That means that the college brand and identity will at least somewhat be tied to them. On the other hand, the blog authors aren’t there just to be a mouthpiece for the college. Sanctioning official blogs means giving up a certain amount of control, which is scary from an administrative perspective.

Interesting times. Once we hash out something I’ll post some kind of summary here (but don’t hold your breath).