Skill Up!


I’ve been spending a bit of time skilling up in work, as I’m settling in and learning the ropes of the new place. As previously discussed, I’ve properly looked at Docker (finally), but I’ve also been improving my knowledge of:

  • Python
  • jQuery
  • Bash scripts
  • NFS

I think my next targets for consideration are:

  • Graph Databases
  • Solr vs. ElasticSearch
  • MapReduce

And here concludeth the most boring blog post I’ve done to date.



Taking Shape


The dream is becoming reality.

Been working hard to get the image sequences flowing nicely over the last week or so, and even I’m surprised at how well it’s coming together. The non-symmetrical aspect is being implemented very nicely (ie. if you triangle with your left leg, an omoplata transition must be with the left leg – and vice versa), all from just one take of each move.

I’m gonna do a bit tonight, possibly implementing the choice of which side to attack (in “neutral” positions where you are free to choose either side), and I will also be looking into processing the background pixels of the move sequences (maybe blurring them or doing something else to slightly reduce their significance).

At some point, I need to refresh my mind on all the remaining tasks before v1.0 is complete, to regain my perspective.

All in all, though, good things are happening.



Dare To Fail

Yesterday I came to the conclusion that I will also need to film the “fail outcomes” of certain moves (which raises the total number of sequences required to 471, at the last count), which instantly raised a few problems with the logistics of how to do this.

For all the other node types where move sequences are valid, it is acceptable to just associate the sequence directly to the unique nodeID.

However, with a fail node, you can have multiple positions to potentially fail into. It is not valid to associate any fail sequence to the resultant position node, since there is always multiple entry points to that move – and, besides, position nodes do not have sequences.

So, the proposed solution is as follows:



User selects “Film Sequence” and then selects the fail node…

    if (noOfOutcomes == 0){
        filmSequence(failNodeID + “.” + positionID + “-“);
    else {
        // render fail node centrally, making use of “selecting fail outcome” boolean, then capturing the selected positionID and making the String for “filmSequence(…)”

Executing Fail Outcome

App randomly chooses the position to fail into.

String for showImageSequence created in same way as before.


So that’s a really basic overview. I will also include some more bits, like rendering a little count of how many positions are associated to a fail node on the positions node in the edit screen, etc. And there I shall leave it for now.



Visual Account Of What’s Left


This is where v0.1 is at, in comparison to where v1.0 will be.

Simple as that – although there should really be about a million percent more columns, all fully green, to represent the actual proportional total of work completed so far.

Not long to go for the finished v1.0, really! But after less than an hour’s work remaining tonight (ironing out a couple of minor bugs), I am done until after New Year. Gonna take a fresh run at it in 2014.

Wish me luck!



Short But Sweet

I need to focus on getting this “example video” produced, so my tasks for tonight are:

  • Do a “welcome” screen from the coach
  • Re-instate flashing nodes
  • Do an “introduction to training” screen from the coach

That should cover all that I need for the first little demo vid. Shouldn’t take too long either.

I will be getting this vid recorded and sent off to the relevant parties ASAP.




A Title Escapes Me

Gonna do a few bits tonight, I’m forcing myself. It’s getting tough to motivate myself at the moment.

My planned tasks tonight are:

  • Film and implement belt promotion sequences
    • Including showing the belt tying sequence upon starting/resuming game
  • Tidy up training screen
  • Try to tie up all loose ends with the advert screens/sequences

I think I need to give this video capture stuff a whirl, too. Need to pull me fingeroot!


EDIT: Just tried the vid capture stuff on the Samsung S3 earlier, it works a treat. Good stuff!