Re: [Nmh-workers] mh-format

[norm]

Ken Hornstein <address@hidden> writes: (14.8 days ago)

>>Who designed it? If I ever knew, I've forgotten. I'm guessing that the answer
>>is that nobody designed it. It probably started out as a simple take-off of
>>printf and grew by small, imperceptible increments from there. I urge that
>>this generation of developers not add any more increments, but invent
>>something better from scratch, or perhaps use or adapt something that already
>>exists.
>
>I was curious, so I looked at this. mh-format did not exist in MH 5.  It
>looks like first public release of MH I can find that contained it was
>MH 6.6 (there may have been an earlier one, but I cannot find it).  So
>probably someone at UCI, likely either Marshall Rose or John Romine,
>wrote it.

It was likely not John Romine. Certainly, Marshall Rose, could have.

>I can't really defend mh-format, other than to suggest that
>once you understand the syntax for branches it's not COMPLETELY terrible.
>It's a special-purpose language designed for a very specific task.  Like
>other languages designed for embedded work (e.g., D) it contains no loop
>operations so it always completes.  Containing only two variables is
>really the thing that sucks about it.

Maybe the first version contained  NO variables, that is, maybe it was
just a simple takeoff from printf.

>
>Lyndon has already mentioned his work on Lua; I think that work is great
>but he also said that it's a lot of work and requires pretty much all
>of current configuration to be chucked over the side.  Since this is a
>volunteer project that's going to take some design and time to get it
>right.  In the meantime we can't really ask people to stop working on
>the current codebase, so we have people working on the next-gen effort
>and improving the current code.
>

    Norman Shapiro