[Oberon] Intermediate scopes in Oberon-07
dsar at eml.cc
Sun Feb 25 13:18:40 CET 2018
On Sun, Feb 25, 2018, at 9:35 AM, Jörg wrote:
> The question is not „fool“ yes or no. Look at NW, he used all three
> techniques in his compilers over time.
> In my point of view, you can see it as kind of evolution in thinking.
This is not evolution, this is cost-benefit trade-off.
There is a level of acceptability and the use of procedure variables is not (for me). But Wirth accepted this trade-off.
Forward reference add a bit of complexity in a single-pass the compiler, that's true, but if you look at OP2 sources it's on an acceptable level. But if you look at Active Oberon compiler where objects defined later in the source text may be used, the parser decisely worsened in readability and increased in size. This is unacceptable on every front, although a very brave work.
More information about the Oberon