[libre-riscv-dev] getting important: need to create the project plan, involves estimating times for tasks (and creating them)

[Luke Kenneth Casson Leighton]

On Fri, Apr 26, 2019 at 11:52 AM Rishabh Jain <rishucoding at gmail.com> wrote:
>
> for pinmux, as earlier luke pointed:
> 1. nmigen version of the pinmux
> 2. Interface Infrastructure Generator

 yes, and even (2) can be subdivided further, into "GPIO header file
generation", "Bus Architecture Generation", and even further into
"individual peripheral infrastructure support" tasks.

> For (1) :
>    > there are many TODO's present in pinmux: many of which are independent
> of bsv.
>     to point some: parser.py needs parameterization, README.md has some
> TODO's
>    >i can start coding in python with a simple case of pinmuxing with only
> 2 peripherals:
>      spi and uart

 yes that would be a great first milestone.

> i guess, it will be better for me to take things step-wise! so, starting
> with writing one module
> at a time, then its corresponding unit test ... similar to soc/TLB.
> these can be my sub-tasks .. and i can put tags like time_to_complete, etc

 better, just put a figure into the "budget".  remember if creating a
sub-task, to:

* make sure that the subtask "blocks" the one it depends on
* make sure that the subtask has "NLnet.2019.02" as its milestone
* *subtract* an amount from the dependent task's budget and *add* that
same amount to the subtask

if you think the task has too low a budget, say so, and we can look at
it.  i set arbitrary amounts.

> luke, what are your thoughts?

sounds good.