Wojtek Walczak wrote:
Max Spevack napisal(a):
>> All your comments are greately appreciated!
>>
> This is all very, very interesting, and I admire and appreciate the time
> that you took to put these ideas together.
>
Thanks, I am glad that you like it :-)
> What you describe is a series of calculations that run on top of that
> data, and that will change over time. I want to make sure that we can
> manage all this data in a way that is well documented and extensible.
>
Agreed. I think that it is a matter of further discussion with Yaakov
to find an agreement about the architectural details of integration
of my ideas into ekg. Nevertheless, I think that following some simple
rules like 'keep it abstract', 'keep it simple' and 'keep it
separated'
should apply.
> Will the algorithms that you propose use the database, or will they
> simply be re-calculated with each run? My guess would be the second,
> because they are based on personal interactions, which will continually
> change.
>
I would say the second too for two reasons:
- we don't need to spoil the way database is constructed
- it's easier to count the results for different date/time slices
if it is dynamically counted
If computations for certain time intervals can be stored in another
table, that would be ok.
For instance, it would save the trouble of coming up with visualizations
that are expensive to calculate -- as they tended
to be before.
> The two primary directions that EKG needs to go are in data
collection
> and data processing/reporting.
>
> Your proposal is certainly in the latter group.
>
Indeed. I still have to assess how much time do I need for all the steps
described before. In case there is some time left, I will think about
gathering the data from git repositories, but I guess GSoC is too short
for all this :-)
Thanks for feedback Max,