Re^12: Is an aXML compiler possible?

[Logicus]

Correct Tense.

I worked on a compiler because that's what Corion suggested I do. That was the reason behind the "too difficult for me" thread where I made it clear that I was struggling to write such a compiler, and did not believe I was capable of doing so alone without help. That was prior to finding out about Plack, and optimising the runtime parser, thus obviating the need for a compiler given the sheer speed of modern server hardware.

That particular wild goose chase cost me a good couple of months development time, and I'm not inclined to persue it further in light of the current speed of the new parser which is well fast enough for my uses.

If after I have proven the value of aXML to the masses, someone feels like creating a compiler for it then they are free to do so, however I will not be bothering as I really have better things to do.

[chromatic]

I worked on a compiler because that's what Corion suggested I do.

The distinction between compilation and interpretation is orthogonal to the ability to write a BNF (or another specification) of the language's syntax. Compilation versus interpretation is a detail of implementation.

What ikegami is trying to understand is if there are legitimate documents you cannot produce from aXML programs because they contain character sequences that could be aXML instructions. (Certainly you can modify the plugin environment to avoid that, but there's a tarpit to avoid in doing so. Whether that's a problem in practice is up to the user to discover, but I've worked around quoting issues in polylingual programs before and have no inclination to do so again.)

[ikegami]

Correct Tense. I worked on a compiler because that's what Corion suggested I do

So which one is it? You didn't bother, or the reason your bothered is that Corion suggested you do it?