|
Post by chriscrawford on Sept 28, 2015 15:09:27 GMT -8
It will be cleaner if I organize this diary by weeks. I don't need to burden the reader with trivial details. My goal for this week is to get the engine components of FroggerLogger operational.
Monday: I have stripped out the Interpreter components and am now focusing on bringing back all the engine components. Just now I am wrestling with the problem that Facundo solved neatly with results that span several entries. For example, when beginning the process of determining whether an actor has a role to play, FroggerLogger adds the entry:
DirObject Camiggdo considering reacting
There follow a series of calculations in the Interpreter, which don't show up yet. If Camiggdo decides to react, then the next line is
Camiggdo executing role Victim
Facundo had come up with a nice way to merge these into a single line, reading like this:
DirObject Camiggdo considering reacting; reacts as Victim
This permits the author to see the result easily without having to open up the nodes underneath it. This kind of problem shows up in many different places in the logger. Should I try to implement it? It will require some serious programming. I haven't yet decided.
|
|
|
Post by chriscrawford on Sept 29, 2015 10:45:11 GMT -8
Tuesday Today I'm just fleshing out more log entries to FroggerLogger. There are a great many such entries, and they must be added carefully, because some entries increase the depth of the tree, requiring me to balance them with "up" commands that bring the focus of logging to the higher level. This is all very tricky, because some nodes could have children, but don't. Some nodes cannot have children. It's all rather tricky business, and if I get the system out of whack, it produces crazy results. Programmers will recognize this as the classic problem of managing a stack of information. If you increment or decrement that stack pointer at the wrong time, you've got big trouble, and it can be hell to track down exactly where you went wrong. So I'm moving very carefully.
Later Well, I've made some progress. I have fleshed out a little. But it is slow.
|
|
|
Post by chriscrawford on Sept 30, 2015 13:20:08 GMT -8
Wednesday My head is too cluttered up with all these technical details. I'm going to use today to catch up on errands and deal with all the other miscellanea of my life.
|
|
|
Post by chriscrawford on Oct 1, 2015 19:37:54 GMT -8
Thursday Another day devoted to my second-tier project, some software for an archaeology professor.
|
|
|
Post by chriscrawford on Oct 2, 2015 19:53:52 GMT -8
Friday I wrapped up this stage of the archaeology project and returned to work on Siboot. The task was to attach values to the nodes in FroggerLogger, a complicated trick that I have been intimidated by. I picked at the problem for a few hours, then went outside and worked on trees. Then I came back inside and fiddled around some more, losing confidence again and taking a nap. Then I came at it a third time, did a little work, and all of a sudden it was working! Programming is really just a matter of typing the right characters in the right places, and if you get them all correct, it's remarkably fast. Of course, we spend most of our time typing in the wrong things, trying them, failing, trying something else, over and over for hours. In this case, I pretty much got everything right the first time. I haven't the faintest idea how I pulled that off...
|
|
|
Post by chriscrawford on Oct 3, 2015 12:31:04 GMT -8
Saturday I think I might have gotten FroggerLogger operational! I'm still checking....
|
|