Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

This should really be two documents #26

Closed
ekr opened this issue Sep 10, 2020 · 14 comments
Closed

This should really be two documents #26

ekr opened this issue Sep 10, 2020 · 14 comments

Comments

@ekr
Copy link

@ekr ekr commented Sep 10, 2020

Despite the way it's phrased as "requirements" but really there's almost no relationship between the use cases where you want to discover which general-use resolver you might use locally (Section 3) and the use cases where you are might discover which resolvers to use for a specific domain (Section 4). It's not at all obvious that you would want to use the same technology for these. If we are to have a requirements document, we should split these into two, one for each workstream.

@ekr
Copy link
Author

@ekr ekr commented Sep 10, 2020

Or maybe 3, because DNS is yet different.

@ekr
Copy link
Author

@ekr ekr commented Sep 10, 2020

In addition, the security properties of these are likely to be very different, in particular because we don't know how to solve the local discovery problem with reasonable security.

@chris-box
Copy link
Contributor

@chris-box chris-box commented Sep 10, 2020

I'm open to splitting it; let's see what further opinions we get from the working group.

@martinthomson
Copy link

@martinthomson martinthomson commented Sep 10, 2020

My suggestion is to cut Section 4. Then the document has just the most important stuff.

@tfpauly
Copy link
Contributor

@tfpauly tfpauly commented Sep 10, 2020

I would prefer to, if we split, split 3 and 4, but keep both. There's mechanism work to do on both, and while 3 might be harder, we shouldn't not do 4.

@ekr
Copy link
Author

@ekr ekr commented Sep 10, 2020

@martinthomson
Copy link

@martinthomson martinthomson commented Sep 10, 2020

4 as a separate document is probably fine. Honestly, most of Section 4 seems weak to me, but 4.2 is a completely different use case that would benefit from standalone treatment.

@Andrew-419
Copy link

@Andrew-419 Andrew-419 commented Sep 13, 2020

I'm struggling to see the advantage of having separate documents versus multiple sections in a single document. I can see advantages in the latter approach, not least of which to have a single repository for all of the use cases and associated requirements, noting that different use cases may share some of the same requirements.

@ekr
Copy link
Author

@ekr ekr commented Sep 13, 2020

I think we should be able to gather consensus on one set of use cases and requirements, while deferring the others. That's facilitated by having separate documents.

@Andrew-419
Copy link

@Andrew-419 Andrew-419 commented Sep 13, 2020

It seems premature to be proposing to split the document without further discussion and does risk some use cases and requirements effectively being abandoned, irrespective of their importance to users.

@ekr
Copy link
Author

@ekr ekr commented Sep 13, 2020

It seems premature to be proposing to split the document without further discussion

I'm proposing it so we can discuss it.

and does risk some use cases and requirements effectively being abandoned, irrespective of their importance to users.

What I'm proposing is that we defer those use cases and requirements until we have addressed the first set of use cases.

@Andrew-419
Copy link

@Andrew-419 Andrew-419 commented Sep 14, 2020

It seems to me that the various types of same provider auto upgrade (ie including with DNS forwarders) are the most pressing need to be addressed. However, it also feels important to identify the other items to include in the use case and requirements universe, even if these are to be addressed later, as well as to be clear which if any are to be excluded.

It would be ideal for the WG to adopt a document that gives clarity on the totality of use cases to be covered, with an order of priority if necessary, rather than to only bring forward a document containing a few at first with an expectation that some others may be done later. So on that basis I'd prefer the document to move as is rather than being split - if there is a view that some use cases should be dropped I think that it would be better to have that discussion now rather than postpone it.

@nicolaileymann
Copy link

@nicolaileymann nicolaileymann commented Sep 15, 2020

I think it's a good idea to start with single document describing all the requirements and use cases which will be covered. If we split the the document now I see the risk that only some of the requirements will be covered and others not (e.g. if only one the documents is moved forward). Therefore I am more in favour of having a single document as a basis and to move this document forward. Lets have the dicussion now what we want to include and what not.

@chris-box
Copy link
Contributor

@chris-box chris-box commented Nov 9, 2020

Good arguments on both sides. I agree we do need all requirements described somewhere, but we also need small enough drafts so that are manageable and can be reasoned about. On the list Glenn proposed dividing into two or perhaps three. I tend to agree, so draft -01 is now a single use case. This does of course mean that other requirements drafts are also required. Authors of those could of course grab text from draft -00 and relevant closed issues in this repo.

@chris-box chris-box closed this Nov 9, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

6 participants