Next week, the last ecography is going to be made and then I'll know, probably, when my 2nd son will be born.
So, until then (and then after, although in a slightly different way... :-)) ), it will be "hard working days".....
There is a lot to be done that I'm really looking forward for the time to do it.
But I think I can't complain, though.
The development work has been in a "slow motion" mode although that is not all bad. At least, the time has been useful to let it gain consistency and ideas become more and more detailed.
Whenever I pass again to implementation time, it will be just to write code and test it. i won't have to be spending time then to seek up how I and to do it.
The Mex4th VTM is almost defined and I needed to make some changes in what was already defined in order to make it more consistent and logical.
From what I can already see from the tests already made, it's code it's going to be quite compact; so, in a carefully made runtime, using the right set of instructions for each primitive, I think I'll be able to achieve a very good processing speed.
And another thing is that most definitions can be just a few instruction long which is very useful to help anyone understand how it is laid out.
And the whole architecture is very much coherent.
The fact of have increased the number of stacks (3 on the VCM - Virtual Code Machine - and 4 on the VTM - Virtual Thread Machine -- the additional one if for the high level structures) does make it necessary to have additional instructions to deal with them (although I'm only with 4 stack handling instructions per stack -- DUP, DROP, OVER and SWAP) but also makes it possible to have almost always 2 elements deep in any stack for any instruction.....
The part that I'll have to deal with next will be the VM-SHARE wordlist, to make it possible to have a lot of cooperating VTMs.
Anyway, soon enough I'll know for certain how everything will go.
Meanwhile, a very good rest of day and good night to you all, here in LJ land!