OSCON : Executive Summary
OSCON is a very different type of show that is dificult to sum up in a few short paragraphs. It’s a unique blend of commercial and open source interests coming together under one extremely pleasant roof, with talks ranging from open source strategy to parody to the point of stand up comedy (albeit with a none to subtle message)
I’ve done a series of posts on the specific sessions I went to and this is an overview of the takeaways and the the face to face conversations I had while there, what I consider to be the highlights of the show for me I guess
- The Strategy
The emerging theme of the show was the need to adapt the successful open source market(movement) to the new world of web2.0, several times it was mentioned that the model where you build code under GPL licence and then distribute works fine when people are actually installing your software on their machines, now that applications are installed on the web the terms have changed, hence Tims oft misunderstood comments about the licences needing a review, it’s not that he doesn’t think they are needed, they just need an update. The story of Flickr and Zoomr pointed out that we also need a new approach to data ownership, I have been thinking this way for some time now with a particular project idea and this was backed up by the creation of a new movement called MoveMyData.org which is being setup to specifically allow you to move your data (sic) between various online services, there was also a call for the creation of an Open Data Agreement and an Open Standards Definition to begin addressing these points, there already exists two related sites on this, Open Knowledge Definition and the Freedom of Data movement, I expect a lot to happen here very quickly, in fact I had dinner on the Friday night with various people, one of whom was r0ml and he handed me a printed card containing a list of compliance criteria for “The open Standards Requirement” which listed the following:
- The standard must include all details necessary for interoperable implementation
- The standard must be freely and publicly available (e.g. from a stable web site) under royalty-free terms
- All patents essential to the implementation of the standard must be licensed under royalty-free terms
- There must not be any requirement for execution of a license agreement, NDA, grant, click-through, or any other form of paperwork to deploy conforming implementations of the standard
- Implementation of the standard must not require any other technology that fails the meet the criteria of these requirements.
You can find more details here
I had a quick conversation with Mark Spencer of Digium after his talk to catch up with what has been happening with Asterisk since ETel where it was everywhere in the show, they have a new offering coming out soon that relates to large CSR organisations, more on that later when we’ve had a chance to catch up. I’ve had several conversations with people about how Asterisk is going to affect our business but I still think they under appreciate how much this project is going to have an affect further down the line. They have recently introduced Bluetooth capability to allow you to seamlessly switch calls when in your home or office for example, not because they had to or that people were asking for it, simply because they could do it very easily and it was fun to do. VoIP is one of those new breeds of technologies that can have immediate and measurable effects on the bottom lines even of the smallest businesses, take the example given by Brian McConnell and setting up a new business in Argentina. I expect a lot more from Asterisk and it’s competitor Freeswitch which can be viewed as Asterisk++ some say
- Asymmetric Competition:
Another popular thread during Tuesday’s O’Reilly Radar sessions, I have to say that I wasn’t sure of exactly what they meant originally but it came to mean the following to me over the course of the next few days: “an asymmetric competition is one in which two opposing forces compete on uneven ground – David and Goliath, perhaps. Think OpenOffice vs Microsoft Word. As pointed out by Nat in one session covering this, it helps to have a lot of money when competing for market share with someone that hasn’t. However just having money doesn’t always mean you will win, sometimes it’s existing market share vs a new entrant of course. The summary seems to be the big 5 vs newcomers.
According to David Skok, one of the original investors in JBoss, the conversion percentage of the middleware firm was around 3%. Put another way, it means that 97% of JBoss users did not pay them anything at all. The question was asked as to whether there was actually enough money in that business model if virtually all your customers decide not to pay, support contracts and maintenance contracts are one way to mitigate this, a customers may pay a lot of money for support and never need it, therefore the cost model is exceptionally good, with a risk that if you make bad software the cost model becomes very high of course. Before acquisition, JBoss were projecting around $65 million for this year, and possibly double that next. Depending on your perspective that could be good or bad, but imagine the profits if the adoption rate was much better. There were several discussions concerning the role of service, support and maintenance, and the general conclusions seemed to be that support itself is not enough. However this is the current most popular route to steady income, but it’s clear that other opportunities need to be realised, perhaps it’s in making software that more widely adopted? Looking at some of the products on offer it’s not hard to see that happening
Tim spoke regularly about Virtualisation and backed it up with some graphs from the jobs market data they are sharing, there is a big rise in this area and a lot of it comes from the growth in colo centres offering virtual hosts to smaller companies I believe (without hard data to back that up though) Expect to see a rise in products that virtualise OS’s for PCs too.
- Early Adopters Might Leave, But Are Replaced:
It was raised in the executive briefing session that while early adopters are attracted to new open source products by their very nature they leave as soon as it starts to become too mainstream, Michael Tiemann from Red Hat suggested that this was not a problem of any kind as the mainstream product attracts greater numbers of mainstream users, he did seem to underplay the ‘power’ that having high profile early adopters use your product can do though. You could also argue that having high profile people adopt your product means that it hits the mainstream that much quicker.
- Open Data:
A final word on open data, as things stand now it is simply too hard for many but the adept alpha geek to move data around, assuming it can be gotten in the first place of course, there are a number of services that now allow you to import your data from an existing competitor, not like the case of Flickr/Zoomr but especially in the RSS feed and OPML file area, in fact any new service in this area HAS to have that option or people simply wont bother to move, once a Blogroll is built you need to be able to easily export to OPML and import to another service in order to compare, you aren’t going to create it from scratch again after all. So in some areas it’s working already, even if the user experience could be much improved in order to appeal to mainstream. The sticking point now is what happens when you have made a large digital investment in a service but want to try out another, at present there are very few options, some new services allow for a suer to import some items from another service, I tried another picture site recently that allowed this (name escapes me now) and the experience was quite slick. However as your investment grows the legacy of all that data, for example in Flickr I have almost 5,000 photos, however even if there were a simple ‘transfer all’ function I would still want all the metadata to go with them, and of course this metadata is worth a lot less when it is taken out of the community that made it, like going to a coffee shop but finding that none of your friends are there… I see some potential in services like Amazons S3 since it could be made so that you always keep your data there but present it to different front end applications?, then the transfer itself would be easy. I’m also sure I heard recently a proposal to embed the metadata in the actual object itself, thereby making it always available with the object, there are big problems here though with authenticity and verification (comment by X in the object is useless if X doesn’t exist in the new service) unless those too are handled by another horizontal platform. Suffice to say its a complex problem that needs considerable thought, a lot of which has been done by a small grpup of people I am fortunate to be part of but thats another post entirely.
In the end I am very glad I had the opportunity to go to OSCON, I met several people outside of sessions, in the ‘hallway track’ as its known and am following up with them on a few relevant threads (Jabber, OSGEO.org, Platial) plus the chance to discuss some of the ‘open’ topics with several new people.
My only real complaint is that there were simply too many seperate tracks going on due to the size of the conference, so it meant making some hard choices between sessions