Page 1 of 1

The RLE documentation is wrong

Posted: May 23rd, 2017, 3:07 am
by blah
The documentation on the RLE format states that values for many-state rules should start with '.' for 0, 'A' for 1, and go up to 'X' for 24. Then they count up from 'pA' for 25 to 'pX' for 48, 'qA'=49, 'qZ'=72. Wait, 'qZ'? If it gets to Z, it would have to be state 74. And this is incongruent with what golly produces:

Code: Select all

x = 256, y = 1, rule = 2/2/256
.ABCDEFGHIJKLMNOPQRSTUVWXpApBpCpDpEpFpGpHpIpJpKpLpMpNpOpPpQpRpSpTpUpV
pWpXqAqBqCqDqEqFqGqHqIqJqKqLqMqNqOqPqQqRqSqTqUqVqWqXrArBrCrDrErFrGrHrI
rJrKrLrMrNrOrPrQrRrSrTrUrVrWrXsAsBsCsDsEsFsGsHsIsJsKsLsMsNsOsPsQsRsSsT
sUsVsWsXtAtBtCtDtEtFtGtHtItJtKtLtMtNtOtPtQtRtStTtUtVtWtXuAuBuCuDuEuFuG
uHuIuJuKuLuMuNuOuPuQuRuSuTuUuVuWuXvAvBvCvDvEvFvGvHvIvJvKvLvMvNvOvPvQvR
vSvTvUvVvWvXwAwBwCwDwEwFwGwHwIwJwKwLwMwNwOwPwQwRwSwTwUwVwWwXxAxBxCxDxE
xFxGxHxIxJxKxLxMxNxOxPxQxRxSxTxUxVxWxXyAyByCyDyEyFyGyHyIyJyKyLyMyNyO!
"qWqXrArB". There is no 'qY' or 'qZ'. That's a simple typo though, I'm more interested in knowing why it only counts up to X instead of Z in the first place. That seems pointless; you're just wasting two good letters (not that we don't already have more than enough combinations for all the states), and why start at 'p'?

Re: The RLE documentation is wrong

Posted: May 23rd, 2017, 4:29 am
by rowett
blah wrote:There is no 'qY' or 'qZ'. That's a simple typo though.
True, thanks for reporting. I've fixed it in the Golly docs.

Re: The RLE documentation is wrong

Posted: May 23rd, 2017, 8:15 am
by dvgrn
blah wrote:I'm more interested in knowing why it only counts up to X instead of Z in the first place. That seems pointless; you're just wasting two good letters (not that we don't already have more than enough combinations for all the states), and why start at 'p'?
Was looking at that recently. I don't have an answer on the decision to start with 'p' instead of 'a' as the old MCell multi-state RLE format does... except that it prevents 'b' or 'o' from being used, which might make it easier for Golly to guess correctly what format a random chunk of text might be in. (Technically you don't get to 'o' if you start with 'a' anyway, though.) MCell had an "#L" prefix to make things pretty clear, but Golly has to work with just the encoding itself.

As far as skipping 'Y' and 'Z', well, MCell format does the same thing for whatever reason. Maybe having the base be a multiple of 8 just looked a little cleaner somehow. A 16-letter subset could have been a contender, but I suppose that would have been just a weird nonstandard hexadecimal.

Re: The RLE documentation is wrong

Posted: May 29th, 2017, 1:03 pm
by wwei23
And when I paste an RLE into Golly from Lifewiki, I get an empty rectangle! Please change it back!

Re: The RLE documentation is wrong

Posted: May 29th, 2017, 1:59 pm
by blah
wwei23 wrote:And when I paste an RLE into Golly from Lifewiki, I get an empty rectangle! Please change it back!
What RLE? It's far more likely you got something wrong.

Re: The RLE documentation is wrong

Posted: May 30th, 2017, 3:41 pm
by fluffykitty
Golly's RLE parser has not changed, only the documentation. Please post the broken RLE here and a link to where you got it.

Re: The RLE documentation is wrong

Posted: June 23rd, 2017, 12:06 pm
by wwei23
There's a space after the rule is stated and it confuses Golly.