Re: [Tails-dev] extension specification

Delete this message

Reply to this message
Author: Giorgio Maone
Date:  
To: sajolida, The Tails public development discussion list
Subject: Re: [Tails-dev] extension specification
On 07/05/2015 18:12, sajolida wrote:
> Hi Giorgio,
>
> A few days ago we finished a wireframe of the extension that we believe
> is a good start for you to work on. It might still change slightly as we
> continue discussing small details of it.
>
> See
> https://labs.riseup.net/code/attachments/download/759/extension-20150430.fodg.

Looks great, thank you!

>
> We also updated the blueprint with some more implementation information.
>
>     https://tails.boum.org/blueprint/bootstrapping/extension/

>
> See the sections:
>
> - Non goals
> - User scenario and wireframe
> - Integration in the web assistant
> - Initial browser detection
> - Checksum verification
> - Data source

Regarding "initial browser detection", I can also provide the JavaScript
for browser sniffing and browser-specific web content provision if needed.

>
> Regarding the data source, we need to know whether YAML works for you as
> we already create very similar YAML files for the automatic upgrades.

It works for me, no problem.
May I just ask, though, why YAML and not, for instance, JSON?

>
> Does all this sounds reasonable to you?

Yes it does.
>
> I think that what we need to work on together now is to specify which
> HTML tools (div, class, etc.) you will need in the code of those web
> pages to be able to do your stuff. Then tchou and I will start writing
> the HTML code for those pages.

I'd actually feel more comfortable with you providing the HTML written
the way you prefer best from a web authoring standpoint, and me
developing my interaction code around the structure you come up with,
bending it to my needs only if really necessary (e.g. by adding a
missing id or class attribute)
>
> If that's relevant for you, note that we will most likely use bootstrap
> for the CSS of those pages. So for example the progress bar would have
> this markup:
>
>     http://getbootstrap.com/components/#progress

That's fine.
>
> Before you can start coding for real we'll also need a name for the
> extension. We'll try to sort this out quickly, see #9295. Feel free to
> comment on the latest proposals.

I cast my vote for "Download and Verify Tails": unambiguous and
future-proof, in case should the verification process change.
>
> In all your work, it's also important to keep in mind that we'll try to
> have a similar extension for Chrome at least very soon. I don't know if
> you can make coding decisions that are more portable than others, but we
> should do that as much as possible.


Yes I can and I had already planned to.
>
> We should also try to make the code as little Tails specific as
> possible. Of course it will be Tails specific as a first implementation,
> but we'd like to keep the Tails specific bits at least easy to factor
> out later on if other distros want to reuse our design and make it more
> generic.


That was my idea too, I don't like to waste potentially reusable stuff.

>
> We also want to ask you whether you think we should have some written
> contract for your work with the specifications, deadlines, conditions,
> and so on. We don't usually do that when we work amongst ourselves but
> if you require one we can maybe get something.

Informal email exchanges like this and the ones you initially contacted
me with will work just fine, thank you.
>
> Otherwise, at least regarding the deadlines we'll have to deliver the
> web assistant to Hivos by the end of the year but we'll do several
> iterations before that and it would be great to have something on which
> we can do user testing let's say before July. Does that sound reasonable
> to you?

I'm currently very busy with the Firefox transition to the Electrolysis
multi-process architecture (enhancing security through low-privilege
content sandboxing), and specifically with adapting NoScript to it and
helping developers of other popular and complex add-ons who are facing
the same challenges.
Therefore, end of July or beginning of August is more realistic IMO,
even though I'll do my best to prototype it ASAP.
On the upside, you're guaranteed of being provided with an
Electrolysis-ready add-on, rather than testing something earlier which
only later is found to cause troubles in multi-process Firefox.
>
> Ah, and we took it for granted but your code will be public, GPL, and
> all from (almost) the beginning, right?

Sure thing!
> Maybe we should create you a
> repo on git-tails.immerda.ch.
>

Why not?
-- G