Out of interest, what is required for a launch window? I assume there's a lot of variables involved.

Launch windows vary depending on the aim of the mission. If it is just a simple earth orbit but non-rendezvous or non-docking type mission, then there is no window at all - apart from maybe hours of daylight windows or weather windows.

If the spacecraft is going to meet up with another spacecraft in earth orbit, such as the ISS,. then the launch window will be primarilly restricted to a period of time within which the ascending spacecraft will achieve an orbit that puts it in the same orbital plane as the target spacecraft.
This will vary depending on the ability of either spacecraft to change their orbital planes once they get into orbit. I would expect that the Dragon will be quite limited in this respect so it will be depending on the initial booster rocket stages to put it where it needs to be in space. This is how the Soyuz system works.

Apollo was very versatile because it had a sod-off (technical term) Service Propulsion System at the back end of the Service Module which allowed it quite a bit of flexibility once it got into space. The upcoming Orion spacecraft will also have similar capabilities. The Shuttle, on the other hand, was a bit limited - although that was as much to do with the sheer weight of the Orbiter as opposed to available manoeuvering rockets.

If a spacecraft is heading out into the Solar System, then the launch windows will be determined by the relative positions of the moon or planets. Moon launch windows occur every month. Planetary launch windows can be years apart.

iirc, there is a very complicated, long winded, and occasionally interestign NASA white paper on statistical analysis of engine parameters during the launch event. With years of launch data NASA can now actual spot an "off trend" event and halt the launch sequence before any serious damage is caused. If i read it right (and it was a while ago) they effectively model the engine and fuel system, and then characterise that model with real measured launch data, effectively "learning" the model for a nominal launch. Any off nominal behavour is then easy to spot !

I'm not sure if Space X would have enough data to do the same thing however?

Due to try again tomorrow morning at 07:44 GMT.

And the BST time?

Watching the countdown now. Let's hope it gets off the ground this morning.

T-10 minutes.

It has a great old fashioned look to it. It reminds me of the older versions of the Atlas.

Well, that was the easy bit - Now we get to see if their on-orbit manoeuvering is up to scratch.

Glad it got into orbit cleanly, pretty epic to be able to watch it live.

Let's hope that this opens the door for more private space firms to keep pushing the boundaries.