11
\$\begingroup\$

This question:

https://codereview.stackexchange.com/q/117311/507

Is more of a design review rather than a code review. I have plenty of valid feedback that could be applied to the question but it's currently marked as off-topic (probably as it stands a valid assessment).

So the question becomes: could it be made on topic (because we can't run it that is not possible as it stands with out current guide lines)? Or is there a case for putting it on-topic because there is a valid review that can be done on it?

\$\endgroup\$
  • \$\begingroup\$ Possible duplicate of Reviewing "design" \$\endgroup\$ – Phrancis Jan 20 '16 at 19:42
  • 1
    \$\begingroup\$ @Phrancis not really.. the duplicate talks about presented fully working code and not about an abstract design. Aside from that: since when do we care what OP wants, right? \$\endgroup\$ – Vogel612 Jan 20 '16 at 19:44
  • \$\begingroup\$ Fair enough, I've retracted my VTC \$\endgroup\$ – Phrancis Jan 20 '16 at 19:45
  • \$\begingroup\$ What criteria do you propose to decide whether a design is complete enough to start reviewing? \$\endgroup\$ – 200_success Jan 20 '16 at 19:52
  • \$\begingroup\$ To be clear — to make a question like this on-topic would require a change or an exception to our rules. \$\endgroup\$ – 200_success Jan 20 '16 at 19:54
  • 1
    \$\begingroup\$ It appears that this question would be too broad for Programmers.SE \$\endgroup\$ – Simon Forsberg Jan 20 '16 at 20:07
6
\$\begingroup\$

As a question, I'd deem it off-topic, too.

But if we consider a grammar definition as reviewable source code, the grammar itself would be reviewable, given that it produces the desired results.

The problem is that it's not asking for a review of the grammar, but OP's question is really:

I'm simply asking for opinions or first impressions. Am I going about this the 'right' way? Do you see any problems with my language defintion, the compiler structure? My plan in general?

If the question can be reworded to be more specifically about the language definition (/grammar), or if it includes the actual compiler source code and becomes a question asking for a peer review of the compiler, then I'd be willing to reopen.

For now it's too opinion-based and high-level design to be a good fit on this site.

\$\endgroup\$
  • \$\begingroup\$ I agree that as it stands it is probably off-topic. What I am really looking for is if design reviews are on-topic and if so what would be the set of criteria we could use to define on-topic for a design review. In this particular case the grammar design is not working code but could be reviewed as a design. \$\endgroup\$ – Martin York Jan 20 '16 at 20:20
  • \$\begingroup\$ @LokiAstari IMO this answer sums it all up already. \$\endgroup\$ – Mathieu Guindon Jan 20 '16 at 20:57
3
\$\begingroup\$

Tl;dr;

Design reviews are on-topic provided there is working code to review.


Take this other question, for example. It's a completely valid CR question because the design is an actual hard-implementation (I use hard pejoratively). There's clear source code that indicates just exactly what the intention was, and the design is completely fleshed out. This means we can go through the code and get an exact feel for how the asker intended the system to work. And this is very good for our community. Several of us learned about a design pattern known as ECS (Entity-Component-System) from an answer (accepted and highest upvoted at the time of this writing) on that question.

The problem with the question in question (https://codereview.stackexchange.com/q/117311/507) is that there is no actual source to review. One of the very first statements in the help/on-topic page:

If you are looking for feedback on a specific working piece of code from your project in the following areas

  • Best practices and design pattern usage
  • ...

The problem is that this question doesn't meet the first requirement (working piece of code).

Should the question be able to be rephrased into a working bit of code, then yes, we can review the code and the design. We cannot do a design review without code to match. If that is what the OP is looking for, then perhaps they should head over to Programmers and read the help/on-topic there. That site is very useful for the higher-level abstraction issues. The, "all I have right now is this thought process, how close am I?" questions.

If all the OP wants is an actual review of the high-level abstraction or design (with or without code), then this is not the place for it, unfortunately.

Disclaimer: the recommendation to Programmers is assuming that the question is reworded to fit within their guidelines. As SimonForsberg said, in it's current form:

It appears that this question would be too broad for Programmers.SE

See: Design review for a compiler question

\$\endgroup\$
1
\$\begingroup\$

Well, as per Programmers SE help page:

If you have a question about...

  • software requirements
  • software architecture and design
  • algorithm and data structure concepts
  • quality assurance and testing
  • development methodologies and processes
  • software configuration management
  • software engineering management
  • software licensing

(emphasize mine). So I'd say that the question as it stand would be perfectly on-topic over there. So rather than trying to somehow find a way to make it on-topic here it looks like a viable candidate for migration.

\$\endgroup\$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .