From:	Elizabeth M Smith	Date:	Tue Feb  5 09:47:45 2008
Subject:	Re: Re: Lang.Net Symposium 2008

On Feb 4, 2008, at 10:14 PM, Wez Furlong wrote:

> Phalanger is largely complete, but doesn't do very much to take  
> advantage of some of the DLR features (particularly with things like  
> eval()).
>
> The mostly-done status of Phalanger means that there aren't many  
> features left to add, so getting to 100% is theoretically faster  
> than doing a from-scratch implementation.
>
> However, the mostly-done status makes it harder and probably a bit  
> less appealing to someone that wants to try their hand at developing  
> a language implementation.  Which is more fun?
>
> - improving the performance of a fast system by 1%
> - improving performance of a hacked together system by 1000%?
>
> I think it's important that whichever route we take is sustainable  
> by the community so that it doesn't die out when a couple of  
> contributors get too busy.  I think that making the initial work  
> easier and more understandable will encourage the community to get  
> more involved early on will create more key people too.
>
> So what we thought was that we could build a DLR implementation by  
> taking pieces like the scanner/parser and the bootstrapping code  
> from Phalanger (which are all really just boiler plate), and then  
> build a new set of guts to target the DLR from there.
>

I'm with Wez in the "let's do a pure dlr implementation" - we can take  
some ideas from phalanger but it's always easier to write shiny new  
code than to try to hack out bad bits of old code.  The sustainable  
thing is always an issue as well, the code needs to be clean enough  
that new people can pick it up rapidly - and we need to get enough  
people working here in the initial phase to allow for sustained  
absences from some developers.

The other thing we need to make very sure of is that the DLR  
implementation acts exactly the same as the PHP C implementation  
(weirdnesses and all) - that means we'll need to be running all those  
tests that have been making their way into PHP CVS.

Thoughts? Opinions? Volunteers?

Elizabeth M Smith