Documetation improvements

[pjf]

If there's documentation for the tagtime tag format, then I can't spot it. Currently I know of:

• Things in round-parens and square parens are ignored.

But I'm sure there might be other things in there. :)

[dreeves]

I think that's it! Real quick for now: UNIXTIME tags (optional comment) [human-readable timestamp] For example: 1234567890 foo (something about foo) bar (something about bar) baz qux [2000-12-25 09:00:00 SUN]

As you know, everything in parens and square brackets is ignored. The idea is that the stuff in parens are comments that should be saved -- it may be valuable annotation of the data. And the stuff in square brackets is entirely superfluous but makes hand-editing the log file easier.

I have my own convention for defining new tags. The first time I use a tag I'll add comment like: 1234567890 wrk dev (dev: development, hacking)

And though I'm sure Paul disagrees, I think it's best to only use 3-letter tags.

[pjf]

pjf@qapla:~/cvs/TagTime$ ./cntpings.pl -v pjf.log | wc -l
1547

I have a lot of tags. I don't think three letters would be enough. ;)

[dreeves]

What do you mean? You're less than 10% of the way to 26^3! (I have almost half as many unique tags as you, btw, and don't feel constrained by 3 letters. I guess I'd actually advocate the principle that the more frequently used a tag is, the shorter it should be. And that anything used daily should be at most 3 letters.)

On Sat, Mar 2, 2013 at 9:16 PM, Paul Fenwick notifications@github.comwrote:

pjf@qapla:~/cvs/TagTime$ ./cntpings.pl -v pjf.log | wc -l 1547

I have a lot of tags. I don't think three letters would be enough. ;)

— Reply to this email directly or view it on GitHubhttps://github.com/dreeves/TagTime/issues/21#issuecomment-14341908 .

http://dreev.es -- search://"Daniel Reeves"

[pjf]

Touché. ;) Longer tag names certainly fit way better into my mind. It's heaps easier for me to write 'dreeves' and 'bethany' than come up with shortenings I'll remember. :)

However this is wonderfully motivating for me to implement tab-completion, aliases, and to dig out my tag hierarchy code. ;)

[pjf]

Apparently bare numbers mean something special to tagtime.

[dreeves]

They do indeed. That's for the tagtime task manager, documented in template.tsk. The idea is to do stochastic evidence-based scheduling [ http://www.joelonsoftware.com/items/2007/10/26.html] by comparing your estimates of how long tasks take to how long they actually take, without, of course, ever having to start or stop a timer.

On Mon, Mar 18, 2013 at 1:15 AM, Paul Fenwick notifications@github.comwrote:

Apparently bare numbers mean something special to tagtime.

— Reply to this email directly or view it on GitHubhttps://github.com/dreeves/TagTime/issues/21#issuecomment-15043190 .

http://dreev.es -- search://"Daniel Reeves" Goal tracking + Commitment contracts == http://beeminder.com