I’ve finally decided to release ActiveDocumentum.  ActiveDocumentum is a Ruby Gem that I created to bring some of the goodness, learned by pulling apart ActiveRecord, to accessing Documentum. Its nowhere near as mature, but I has been doing the job pretty well so far for the scripts and sites I’ve been using it for. It has a dependency on JRuby because it hooks into the DFS client libraries to facilitate connecting to Documentum. I’m going to post some further examples, but for the time being, here it is.

http://www.github.com/vertis/active_documentum

You’ll need a Documentum repository, JRuby and a copy of the DFS sdk to play with it.

Update: I’ve also done up a quick sample which you can find on my github (right next door to the actual library).

http://www.github.com/vertis/active_documentum_sample

Myth has it that ‘Kilroy was here’ was the mark of a shipbuilder before/during WWII, and that it was picked up by Servicemen who found it amusing that the mark was made in supposedly unreachable places. They then spread the mark elsewhere, as locations around the world were liberated.  I love this concept, because I’m not a fan of limitations — I’m an avid gamer, but I rail against games that don’t let you stray from the path. There is no variety, or replay value because each time you have exactly the same sequence of events.

Many hopeful web developers fall into this pitfall as well. I was reading http://www.adaptivepath.com/blog/2009/10/27/desire-lines-the-metaphor-that-keeps-on-giving/, a(nother) look at a metaphor called ‘desire lines’, that has similarly important lessons for the web2.0 community. You can’t force people to do what you want. The lesson is simple and played out all around us, you can set down a path, but you can’t make people want to use that route. A good designer will probably preempt a lot of what people will want, but not everything, and the overnight success that all of us are searching for is a bit hit and miss. Plenty of good ideas are out there, but finding something that people actually want, and then getting the word out there that you’re product is there is tough.

Or something like that.

Apparently the idea for my Google Maps API part of my challenge has already been done by foursquare. Dammit, that was a good idea too. They’ve taken it to the next level as well, making it possible to complete badges (much like Stackoverflow and/or Kongregate). It’s a bit limited, in that you can’t have it in every city right now, but it should be fun to play with and might get a solid following from people with smartphones and people that love the whole geo-tagging phenomenon. Oh well, back to the drawing board, this was probably a bit complex for a week long project anyway.

One of my favorite finds this year is Programmable Web, a site dedicated to APIs and Mashups that are the lifeblood of web 2.0. Having just watched the movie Julie & Julia, which I highly recommend, I think that its only appropriate that I set myself a similar challenge. Take the top 50 apis, and within the next year develop a non trivial application for each of them (in no particular order). I’ve set myself a couple of rules:

• You are allowed to use any client libraries if they exist.
• Using multiple APIs in the one application is allowed (yay for mashup’s)
• Any libraries/applications that are developed in the course of the challenge must have tests/specs and documentation.
• And of course blog about each api.

First up:

• API: Google Maps
• Deadline: Thursday the 29th of October

You’ll be hearing from me soon. In the meantime I welcome anyone that wants to participate. Just leave a comment on this post, with your details. If it proves to be popular then I’ll get some kind of signup/participant website going.

Spent most of the night downloading Documentum 6.5 components for Linux & Oracle. While I would prefer to use an open source database like PostgreSQL as the backend, the actual database isn’t going to matter much once its installed (they don’t offer it as an alternative anyway). What I’m looking to do is have a D6.5 environment that I can tune to be as fast as possible. This includes doing things that you wouldn’t be able to do on a System that has to be supported, such as using Nginx as the front end instead of Apache Httpd  with Apache Tomcat.

I’m still a bit up in the air as to which App Server I’m going to use. Tomcat is the obvious answer and the starting point for clients that don’t want to invest in IBM Websphere,  BEA(Oracle) Weblogic, or Jboss. I do think that Apache Geronimo and Jetty are worth investigating as alternatives. One of the complaints that I’ve heard against Tomcat in production environments is that it is ridiculously unstable when running Webtop. With the app servers needing to be restarted every week (or more frequently). IMO there is more likely an issue with Webtop itself rather than Tomcat. I haven’t heard of stability complaints when running other webapps.

One of the big complaints against Apache Tomcat is that it isn’t a full stack App Server like IBM Websphere, for instance architecturally IBM Websphere has a lot more under the covers. I’ve not installed Apache Geronimo before but it’s supposed to be more of a complete App Server. I’ll let you know how my project goes anyway,  got to run off to work now.

I’m not a graphic designer. But my use of Twitter has done exactly what I’d hoped and been a focus for my thoughts. As an individual if I want to pump out an impressive web site I have to be able to do every part of it. Unfortunately that means being able to implement a CSS design.

Admittedly you can rely on sites like OSWD, OpenDesigns and OpenWebDesign for a while, or you could buy a commercial design. There is also some good *base* templates around, but in the end though you’ll need to customise those designs at the very least, and at the very worst you won’t find something you need.

With this in mind I decided that it was high time that I created some simple designs of my own, using the best techniques I can manage. They won’t be perfect, but they’ll be a starting point. I’ll post my first results in a couple of days. In the meantime I’m very interested to talk to people that do Web 2.0 designs and pick their brains as to the right way of doing things.

I’m very fond of Eclipse. Sadly the code I’m working on at the moment is mostly C# .net (with at tiny bit of VB), that means I have to adapt to working with Visual Studio. Visual Studio has come a long way since I last used it but I still miss the power that you take for granted when you use eclipse. So I was wondering whether anyone had experimented with the couple of C# plugins for eclipse with anything more than hello world applications.

I’ve been doing quite a bit of work to get some tools for etsy back online. The original ‘etsytools’ was forced to close after my girlfriend recieved death threats, not that they could be taken seriously, but it still put a black cloud above an effort that was primarily designed to benefit etsyians. Enough time has passed that I now feel comfortable reviving the site. It won’t come back in the same form, if anything I’ve spent the last six months building tools that will be far more useful than anything I previouslyoffered. As soon as there is a moment to spare from my day job (which keeps me pretty damn busy), I’ll polish off the last of the tasks required to get it ready for use.

It’s rather expensive to get into iPhone development. For a start you need Mac OS X 10.5 and therefore a Mac to do development. It doesn’t surprise me that Apple on
ly offers the SDK for OS X, but it is a bit disappointing. Its for reasons like this that I can’t wait to see the proliferation of open hardware.

I upgraded my rails sites to use Passenger aka Mod Rails this week. I’d like to say its been all roses, but the truth is that my Typo blog would swear under oath that its not the case. I was quite happy with the out of box Typo + Sqlite combination, and that worked fine under mongrel. Not so under passenger, the website will work perfectly, but I won’t be able to get into the admin section. As you can see I managed to get it working, I switched the database to mysql and its been working fine ever since. I’ll get around to working out WHY it doesn’t work with sqlite at some point (and get around to extracting my lost blog posts).

Other than that small hiccup I’ve been extremely impressed with Passenger. Its certainly a step forward as far a simple rails production hosting. I’m not saying that its time to throw mongrel out the window. But when it comes to trouble free hosting, mongrel is just not there. I want a solution where you can do a simple config and forget about the site. Managing multiple clusters for each of the various sites is something that makes it harder to whip up simple sites.

Passenger allows the user to to configure an apache virtual host with a couple of rails specific parameters (more if you want to take advantage of somecool feature, or do abnormal things), and restart apache and you’re all up and good. Restarting your rails app can be accomplished by running ‘touch tmp/restart.txt’ in your rails root. You’ve then got a very simple solution that will allow you to host multiple sites without doing clustering.