Re: [Snowy] Snowy Hackfest Proposal



On Wed, Aug 25, 2010 at 2:45 PM, Jeff Schroeder <jschroeder gnome org> wrote:
> On Wed, Aug 25, 2010 at 1:08 PM, Sandy Armstrong
> <sanfordarmstrong gmail com> wrote:
>> On Wed, Aug 25, 2010 at 10:02 AM, Jeff Schroeder <jschroeder gnome org> wrote:
>>> Releasing Tomboy Online 1.0 requires:
>>>    - Some css / template love
>>>    - Banging out any openid bugs that we see in the error logs
>>>    - Properly fixing the innodb issues as snowy requires working transactions
>>>
>>> I was thinking of initially making snowy a "read only tomboy note
>>> syncing / viewing / sharing platform". That keeps the feature set
>>> small and makes the release goal very attainable. Scope creep will
>>> bite us hard if we add too many ponies at first. After beta (and maybe
>>> pre-1.0) we can work on enabling the funcooker by default and editing
>>> notes from the webui, encryption, etc.
>>
>> I completely agree with you, except that I don't think we have any
>> hackfest-warranting issues that are blocking a public beta.
>>
>> The way I see it, we're only a couple of patches away from an
>> invite-only alpha, and the only thing *really* blocking a public beta
>> is confidence in our ability to handle so many users and their data.
>>
>> If we're going to have a hackfest, we should be working towards some
>> big stuff (ponies), otherwise it's a waste of the Foundation's money
>> to get together.  If we're not confident that we can work on some
>> ponies in October, we should aim for a later date for the hackfest.
>
> You are right. What major things do we want to work on for the hackfest then?
>
> What is stopping us from enabling the funcooker other than server side
> editing not being supported. If thats it, I could work on editing. It
> can't be much harder than posting the new data, updating NoteInstance,
> and updating the user's sync guid more or less.

I believe the main issue Brad ran into is that every browser's
content-editable generates different HTML, making reliable transforms
to Tomboy note XML challenging.  And popular mobile browsers like the
iPhone (at the time, at least) didn't even support content-editable,
making in-browser editing less useful of a feature.

It could be that existing JS editors have more reliable HTML output
than content-editable via funcooker, but I really have no idea.

Performing the sync itself is, as you say, not particularly difficult.

> Scaling django is just like scaling any other webapp. Measure,
> optimize, measure, optimize, repeat. It works for the mahalo.com guys
> (right down the street from here) so it will work for us. Lets add
> investigating the django caching layer + memcached to "the list". Can
> someone open a bug for that? I'd be interested in working on that as
> well.

Good idea.

> Finally, I tested (by changing them manually in the admin) assigning
> multiple openid urls to a single user. Snowy supported it all
> perfectly. I was thinking of adding another instance of the
> openid-selector in the preferences page for doing just that. More
> ideas welcome.

Very cool, that would rock!

Sandy


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]