Hi Yichuan Duan, I've taken another look; many of the same comments from last time still apply - especially those related to time estimates.
When we say we want a detailed plan, we mean *detailed*. *Nobody* - I don't care how long they've been working with software - *Nobody* can estimate "3 weeks" for anything and be remotely accurate. In my experience, the accuracy of any estimate greater than 2-3 days is usually "a complete guess". Building an estimate at this level of granularity takes time - but it's also part of making sure you actually understand the problem. Historically, we've preferred estimates that are mostly at the level of 1 week granularity; 2 weeks blocks are *occasionally* allowed for a particularly complex piece of work. However, that 2 week block must then be accompanied by a *detailed* description of exactly what is going to be done. To pick on one particular estimation task as an example: 5.2 "write backend specific tests". Why is this a 2 week task? That's 10 working days, or 80 working hours (assuming a standard work week). Lets say that it takes 10 minutes to write an individual test case - that means you're proposing to add 480 new tests to the Django test suite. Where is the inspiration for these new tests coming from? There's only 195 open tickets on database backends - and you won't be able to use those tickets for inspiration, because those problems haven't been fixed yet. Even if the same test applies to 5 different backends, you're still looking at almost 100 unique tests. Looking at that math, your estimate of 2 weeks to "write tests" looks *incredibly* padded. That said, you may have something in mind. You may have ideas about what you're going to test, and why. However, none of this is communicated by your proposal. The process of estimation isn't to write down a series of bullet points that adds up to 12 weeks. It's to convince us that you have enough work to fill 12 weeks. And, to be clear - we already believe that there is 12 weeks worth of work here - the problem is that you need to convince us that *you* know where those 12 weeks of work are. Yours Russ Magee %-) On Sun, Mar 22, 2015 at 8:32 PM, Yichun Duan <[email protected]> wrote: > Thanks. Updated. :) > > 在 2015年3月18日星期三 UTC+8上午6:39:46，Russell Keith-Magee写道： >> >> Hi Yichun Duan, >> >> Thanks for submitting a proposal - what you've submitted is a good start, >> but it needs some more detail and clarification. >> >> 4.1: You say you're going to save the database-specific tests "to another >> place" - Can you give specifics, or at least an idea of you current >> thinking? How will that chosen location allow external database backends to >> provide tests? >> >> 4.1 (the second one): What backend specific tests are you proposing to >> write? We've got a fairly well tested code base; if there's a specific area >> where you think more tests are required, you should be explicit about the >> nature of those tests. >> >> 4.2: Isn't really true. We already have database specific tests, using >> the @skipUnless(connection.vendor == '...') decoration on tests. What were >> you thinking of introducing in this step? >> >> 4.3: If I understand this, You're allocating a week to *run* a test >> suite? This sounds... very well padded. Unless you've got something >> specific in mind, I would consider "running the test suite" to be just part >> of normal development, not something that requires a specific time >> allocation. >> >> 4.4: "Solve the database problem". Sorry, but this sounds like an >> "underpants gnome" answer: Step one - steal the underpants, Step two...... >> Step three Profit! It's not clear to this proposal that you've described a >> single problem, let alone proposed a solution that can clearly be >> implemented in 2 weeks. >> >> 4.5: What constitutes a "more complex environment"? You either have a >> MySQL database, or you don't. What "complexity" are you referring to? >> >> 4.6-4.8 are broadly Ok - having a couple of weeks for cleanup and >> extension tasks at the end of your project is a good idea. However, you've >> allocated 1/3 of your project time to "cleanup" tasks. That seems a little >> excessive. >> >> Item 4.4 is the biggest issue in my opinion. It feels like this is the >> real "meat" in your proposal, but it is not clear to me at all what you're >> planning to do here. You don't have to provide detailed specifics, and what >> you propose here may not end up being what is accepted -- but some broad >> architectural ideas would be very helpful. >> >> It would also be helpful to think about what you can provide as interim >> deliverables for this project. Ideally, the result of this project wouldn't >> be "one big PR at the end" - it would be much better to see a number of >> smaller pull requests. Smaller PRs are more likely to get merged, and >> reduce the need for any time allocation for merge tasks. >> >> Yours, >> Russ Magee %-) >> >> On Tue, Mar 17, 2015 at 8:31 PM, Yichun Duan <[email protected]> wrote: >> >>> I've written a draft proposal according to your idea. draft proposal >>> <https://github.com/friyal/proposal/blob/master/yichun-proposal-draft.md> >>> I'm looking forward for your feedback. Thank you for your help. >>> >>> 在 2015年3月16日星期一 UTC+8下午7:22:40，Shai Berger写道： >>> >>>> I can't answer for sure because there are still open issues there (e.g. >>>> what >>>> do you do in the face of multiple databases using different engines). >>>> But for >>>> the "clear cut" issues, my estimate is that they shouldn't take more >>>> than a >>>> few days. >>>> >>>> Shai. >>>> >>>> On Sunday 15 March 2015 14:03:44 Yichun Duan wrote: >>>> > And another question, how long will this task take in your opinion? I >>>> can >>>> > hardly estimate it now. >>>> > >>>> > 在 2015年3月15日星期日 UTC+8下午7:07:11，Shai Berger写道： >>>> > >>>> > > On Sunday 15 March 2015 03:45:16 Yichun Duan wrote: >>>> > > > Thank you. When you say "backend-provided tests", do you mean >>>> MySQL >>>> > > > scripts, Oracle scripts or something else like these? Or just >>>> some >>>> > > > tests based on django standard, which have marks to tell test >>>> runner >>>> > > > the >>>> > > >>>> > > backend >>>> > > >>>> > > > they suite? >>>> > > >>>> > > I mean standard Django tests, which are located in folders inside >>>> the >>>> > > backend >>>> > > source. That is, Oracle-specific tests in django.db.backends.oracle. >>>> tests >>>> > > etc. >>>> > > Today, the test-runner will not discover such tests and so will not >>>> run >>>> > > them >>>> > > even if the specific backend is in use. >>>> > > >>>> > > In the current Django test-suite, there are several tests marked to >>>> run >>>> > > only >>>> > > on a specific backend. I'd also like to remove them from there, and >>>> put >>>> > > them >>>> > > into the backends. That's the "clean-up". >>>> > > >>>> > > Shai. >>>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "Django developers (Contributions to Django itself)" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> To post to this group, send email to [email protected]. >>> Visit this group at http://groups.google.com/group/django-developers. >>> To view this discussion on the web visit https://groups.google.com/d/ >>> msgid/django-developers/1f8f91d1-e921-4dc7-9ac6- >>> efc61f667660%40googlegroups.com >>> <https://groups.google.com/d/msgid/django-developers/1f8f91d1-e921-4dc7-9ac6-efc61f667660%40googlegroups.com?utm_medium=email&utm_source=footer> >>> . >>> >>> For more options, visit https://groups.google.com/d/optout. >>> >> >> -- > You received this message because you are subscribed to the Google Groups > "Django developers (Contributions to Django itself)" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at http://groups.google.com/group/django-developers. > To view this discussion on the web visit > https://groups.google.com/d/msgid/django-developers/1bee0188-29fe-41d9-bce8-5f5695f683d7%40googlegroups.com > <https://groups.google.com/d/msgid/django-developers/1bee0188-29fe-41d9-bce8-5f5695f683d7%40googlegroups.com?utm_medium=email&utm_source=footer> > . > > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Django developers (Contributions to Django itself)" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/django-developers. To view this discussion on the web visit https://groups.google.com/d/msgid/django-developers/CAJxq84_Jy9uynOawWwxzPCgQ3f2WtAW%2BWyJ08HkSTjQqbAzTow%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
