← Back to team overview

elementary-dev-community team mailing list archive

Re: The future of appcenter

 

Hi guys,

I am not 100 % favorable to a fully online AppCenter and let me explain why.

Populating an online database with all the infos of all the applications in the repositories would be an enormous waste of time and it should be updated each time a new thing is added to any repo. Moreover having everything online would make AppCenter slower while getting package's infos.

Payment apps should be the only apps of which details should be online because of we can't have them in the repos.

What I want online are the following features:

1) Screenshots: we need to abandon debian.screenshot.net platform in order to have screenshots taken in elementary OS.

2) Ratings & Reviews: Ubuntu servers seems to have read only permissions but, even if I am sure that investigating a bit we can obtain write permissions too, it's better to implement our own R&R platform. That's because we want to have a "full elementary experience" i.e. some elementary applications could look bad with Ubuntu's theme and we don't want people complaining for the bad aspect and giving us bad feedbacks.

3) Banners and/or home page: about this point I am not sure. Having banners download by AppCenter would be quite easy but maybe a whole home page could be a better idea.

Before we discuss what technology to use we have to discuss how we want to structure the platform itself.

Maybe having a developer meeting as we used to do would be a good solution to define all these points. 

Best regards,
Mario

Il mar, mar 26, 2013 at 3:17 ,Goncalo Margalho <g@xxxxxxxxxxxxx> ha scritto:
I'm very calm, don't worry, what I would like is to discuss that question "What do you guys want from AppCenter?" I see and app center like the apple AppStore or the google Play. what about the others?

Yes, having code to talk about it's good, when you have 200 developers you ask 10 to create something and then compare. but at the moment there are just few developers, so let's define, plan, discuss and when we know what we want to achieve let's try to create it.




On Tue, Mar 26, 2013 at 1:48 PM, Raphael Isemann <teemperor@xxxxxxxxx> wrote:
@Goncalo Margalho, calm down. There is no decision made at the moment or anything like that. I also don't think that Strobl wanted to nuke your part of the discussion with a "i'll make some proof-of-concept"-code. If we have code to talk about, we just have a better foundation for a discussion, don't mind him.

Before you go crazy here with that stuff, you should first answer that:

What do you guys want from AppCenter? A user opens it and then he sees the newest/hottest/whatever apps with some infos (my guess)? Define what exactly is your target when proposing a radical change to a project (So far no one has answered shnatsel's question about the problem you see in the current AppCenter).

Someone want a webpage instead of AppCenter (if i read that correctly)? Why? 

You should first answer that before going into technical details.

2013/3/26 Chris Timberlake <game64@xxxxxxxxx>
The first decision needs to be "What is AppCenter?" Is it going to be an App Store or is it going to be just an App Center as it is now? That decision changes the course of the project forever. An AppStore is a huge undertaking that should be planned now.


On Tue, Mar 26, 2013 at 5:51 AM, Joshua Strobl <truthfromlies@xxxxxxxxx> wrote:
As I said, these decisions need to be a group consensus. We're not going to be the next Canonical.


On Tue, Mar 26, 2013 at 5:49 AM, Goncalo Margalho <g@xxxxxxxxxxxxx> wrote:
If you do everything on your own deciding everything, there's no need to follow that. 
Saying that. Good luck with the AppCenter ;)


On Tue, Mar 26, 2013 at 12:43 PM, Joshua Strobl <truthfromlies@xxxxxxxxx> wrote:
There is a limiting factor on what we can implement from third parties. For instance, with Ubuntu Reviews API, we (along with everyone else) has read-only access, therefore we are not able to apply our own ratings and reviews (obviously a write process). This is already going to be covered in the API, I'll be pushing out code by the end of the week (hopefully) that will handle a portion of this.

The general idea is to either completely pull all the reviews / ratings from Ubuntu, pretty much regarding every application (although I'd prefer we only limit to applications that are actually popular) and store them in our own database. This will ensure that any breaking changes that occur in Ubuntu's Reviews API do not affect AppCenter, since the reviews are stored with us anyways. Another idea would be to continue pulling reviews / ratings from Ubuntu's Reviews API and only store reviews / ratings by elementary OS users.

It is really up to group consensus. This isn't so much about rewriting things, its more like leveraging existing APIs to get a good jumpstart on an AppCenter.

I would appreciate if you'd follow https://bugs.launchpad.net/appcenter/+bug/1091406, as I'll be posting details, potentially initial JSON formatted string files (for showing how some of the data will be structured when being requested via an HTTP Request) and at some point I'll link to the repo for the API.

- Joshua Strobl

On Tue, Mar 26, 2013 at 3:29 AM, Goncalo Margalho <g@xxxxxxxxxxxxx> wrote:
So we are going to rewrite it? Why in linux community people like to rewrite things? We need to plan stuff to work on in and then implement. Here, everyone likes just to implement. Why dont we think about the future. Use our brains to build something that it will stay like this?

On Mar 26, 2013 10:07 AM, "Sergey "Shnatsel" Davidoff" <sergey@xxxxxxxxxxxxxxxx> wrote:
2013/3/26 Goncalo Margalho <g@xxxxxxxxxxxxx>
I think that the AppCenter now is just a wrapper of packagekit, i mean, instead of using apt you use AppCenter, how do you add reviews? paying apps etc?


No, it's not. PackageKit API does not provide application screenshots, for example. They're fetched on-demand from http://screenshots.ubuntu.com/ or http://screenshots.debian.org/ (they're the same website anyway).

As for paid apps, there's a staggering number of possibilities. Ideally we'd use something distribution- and vendor-independent, and I have a few ideas on how to achieve that. But IMO it's too early to discuss implementing paid apps yet. We'll design the architecture for that when we get there.
 
--
Sergey "Shnatsel" Davidoff
OS architect @ elementary





--
Mailing list: https://launchpad.net/~elementary-dev-community
Post to     : elementary-dev-community@xxxxxxxxxxxxxxxxxxx
Unsubscribe : https://launchpad.net/~elementary-dev-community
More help   : https://help.launchpad.net/ListHelp





--
------------------------------------------------------------
Chris Timberlake
Technical Architect
Phone: 515-707-5109
Game64@xxxxxxxxx

--
Mailing list: https://launchpad.net/~elementary-dev-community
Post to     : elementary-dev-community@xxxxxxxxxxxxxxxxxxx
Unsubscribe : https://launchpad.net/~elementary-dev-community
More help   : https://help.launchpad.net/ListHelp




--
Mailing list: https://launchpad.net/~elementary-dev-community
Post to     : elementary-dev-community@xxxxxxxxxxxxxxxxxxx
Unsubscribe : https://launchpad.net/~elementary-dev-community
More help   : https://help.launchpad.net/ListHelp





Follow ups

References