Thursday, April 17, 2008

Yet mooooore Goooogle

...a final question for Reid was about LinkedIn's support for the OpenSocial API's that are co-ordinated by Google. Something else to contend with I suppose, at least this is an "Industry" initiative and not just Gooooooogle again.

Liferay have been thinking about this so we may not have to invent too much new stuff ourselves hopefully.

Linked In?

I recently set up an ID on LinkedIn after someone I know invited me to join it, then people I used to work with invited me to join their "network" too, then a couple of recruitment agencies I used used to recruit asked me to join their networks, then a couple of recruitment agents contacted me to suggest a few jobs I might want to apply for, then I heard on Radio 4 "In Business" that some companies were checking up on people via LinkedIn - i.e. they were contacting people in candidates networks and informally asking for a reference. At that point I decided to add some useful information there. I've now set up an Alumni Group for a company I used to work for and invited a few ex-colleagues to join and now have 17 people in the group after only 5 days. Several people are those I didn't actually know, its interesting how the word has spread about it.

Having spent more time on LinkedIn recently I have been impressed with how clever it is - the functionality is quite neat the way it presents you with information about your network and their activities, etc. It's also very quick - to explore how they do this I did a search on Google and as usual got distracted with some background to the company behind it. The key founder is a guy called Reid Hoffman and at the Wikpedia link there's a link to a talk he gave recently on "his thoughts on launching and growing a successful technology business". This covered some stuff directly relevant to MediBlog, one of the immediate messages that hit home was that he was not too interested in the actual revenue streams, instead the focus of a start up should be to have ideas about "placeholders" for your advertising or other revenue ideas plus:

1) How do you get your first million users?
2) How do you get to ten million users?

That is quite scary, the implication being that a start-up should plan to have capacity for 1-10 million users without necessarily have all the revenue to support it. That will require some confidence when we have to move off the initial set of hardware - we probably need to think about that one day soon.

Thursday, April 10, 2008

Scalable Social Networks

Brian Chan, one of the Liferay founders is presenting at Jax 08 about how to use Liferay for scalability:

"How do I scale massively the way Facebook and MySpace do? Both Facebook and MySpace are written in scripting languages that have inherent benefits that traditional Java technology-based portals lack. The session shows how to run your own Portal without using HTTP sessions, so that a deployer can scale the same way Facebook scales with PHP and MySpace scales with ColdFusion."

Of course, will be outrageously successful and all our problems will be related to scaling up to meet the demand. Well possibly. Seriously though, it would be interesting to hear this talk as I have some concerns about Java scalability vs performance following uncertain experiences with SAP Web Application Servers, sadly time & money are against a trip to Germany to hear it :(


I had a brief conversation with Steve yesterday afternoon as he'd just got Liferay up & running on Glassfish & MySQL and wondered what best to do next (clearly there must have been some downtime between Jerry Springer and pub opening times...) . Last year I had an estimate from Liferay to configure the email portlet to work on the site and was quoted 3 days, on that basis one would assume that its a difficult task so I suggested "get the email portlet working" as the next step - saying "respect due" if he managed to do it.

A few hours later got the text "You may respect me at your leisure, mail portlet up and running".

I wonder how he did it..?

Wednesday, April 09, 2008

...and on...

Update so far: after some thoughts of offshoring the development to India it turns out that Steve has returned to the UK for a while and has some time available to do some more work on the project. Hurrah.

Here's what's happening now:

* All (re)approved the idea of activating the company; to confirm, that means we all have 25% each
* Everyone is happy to invest the required cash to get the proof-of-concept on the net.
* We need to have some mechanism in place for recognising effort - i.e. if for some reason one of us can't contribute time and/or money then we can't "Carry" them forward indefinitely and so that person's shareholding must be diluted somehow. Similarly, if someone is carrying the whole project for long periods of time then that also needs to be recognised somehow.
* we need something in the M&A about the existing shareholders getting first refusal on sale of shares.
* Offshore development: as Steve is offering to spend time getting the Portal on the net then we'll postpone that.

Everyone: pitch in ideas - technical, business, branding, anything and everything
Jeremy: Company secretary, marketing and business development, "accountant"
Mike: DBA, database design, java development, reporting
Richard: CEO, strategy, technical development: java development and portal admin, domain registrar admin
Steve: CTO, system admin, development framework advisor, java and portal development

Next Steps
Jeremy & I are meeting next week to hand over company papers so he can activate it; Steve is figuring out the portal implementation issues - especially the Glassfish/MySQL installation and configuration of the standard email and chat portlets, we will all write cheques or transfer cash electronically to the new bank account and then we'll procure the server so that I can start to change the domain DNSs, etc. and Steve can deploy the portal. Mike will have some time in around June/July to produce the database design for the first application and populate it with the pre-requisite reference data. Then we will have the hard graft to build the java applications.