A namespace notation whose name is unresolved (e.g. no label) refers to the unique special unnamed namespace that corresponds to the number of dots. Instead of calling these "special unnamed namespaces" we could call them namespace 1, 2, 3, ... based on the number of dots.

E.g.:
. refers to (special) namespace 1
.. refers to (special) namespace 2
... refers to (special) namespace 3
and so forth

That might conflict with actual namespace 1,2,3 though.

Terminology is difficult when the language has such flexibility.

Can a non-special namespace have a name that starts with a number?

Documentation:


Is your example an exception to the concept (rule) that a name does not begin with a decimal digit?

Perhaps the '?' suppresses the interpretation as a number.

This is true. In a larger context the namespace can be used without the question mark. ... quite useful feature, tbh.

I appreciate the example. Then, with regard to the point of my OP, one could call the namespaces:

. Special Namespace 1
.. Special Namespace 2
... Special Namespace 3
and so forth

"unnamed" seems more literal to me. Also, I always imagine there could be other "special" namespaces in the future, and it reduces the need to re-learn old concepts with new names. The language certainly has corners. Maybe different naming would see these special namespaces with more use?

I just think the manual is very dense and could probably be expand to a 500 page manual with extensive examples. Only so many hours in a day.

It's a suggestion. No need to rush a decision. As a very recent (new) student of FASM, I can offer an objective experience/perspective. I.e. I understood the concept, but I think that a specific moniker is helpful to me when I encounter documentation of syntax that corresponds to a uniquely identified entity, as is the case for the specific namespace that corresponds to a specific number of dots.

The goal is to be helpful, of course. FASM appears to be a powerful tool, so it's worthy of large growth/uptake IMO. Granted, assembly is not a popular language in general, but I think FASM could even change that e.g. grow as a tool for higher-level language development as an abstraction layer above the machine. fasmg appears to be geared toward this.

(I have an assembly/C/C++/C#/F#/JavaScript/HTML/CSS background and am interested in memory safety language features for C as well as a language with modern syntax. I'm excited to have found fasm/fasmg. I have encountered fasm, e.g. in web searches over the years, but only recently made the decision to use assembly language and learn and use fasm.)

Ironically, even your exposition that a number could be used as a namespace name by use of preceding '?' could perhaps be clearer in the documentation:

So, neither a label nor a variable can ever begin with a decimal digit. These entities and other types of named entities, such as namespaces could be introduced along with which ones the '?' is applicable to. I think that, e.g., a table of all named entity kinds and a column for whether they can ever begin with a decimal digit and a column for whether '?' can override and what '?' can override, would help delineate the various possibilities.

I'm sorry if it seemed like I was down-playing your feedback - that is certainly not my intent.

The nuance you've reflected on regarding '?' might arises because it was added later - it becomes an exception to prior statements. Hence, why it concludes section two: Perhaps that last paragraph could be worded differently to make it clear a number is prepended with a '?' becomes an identifier.

No problem at all, and thank you. You entertained my feedback and shared your thoughts. Discussing, itself is a good exercise for my learning. I can see that fasmg has rich capabilities and I have much to learn.

I added a small example. My preferred solution is to give more examples, thus I've been adding more of them recently.

Sounds great. Agree, examples are the crestal explanation.