276 replies to this topic

[craigix]

A little summary of the categories:

I don't want this thing like a ghost town, remember the Pandora scene is likely to be small first and grow in a smooth way over the first year, if you look at the gp2x archive some of the categories are almost empty.

What I have done is have minimal categories in each section, but have a very good search function.

It's quicker to type 'mame' in to the search box (rather than doing: 'emulators' 'arcade' 'mame'), because the search will know you want the latest MAME and will show it with the download link before it shows anything else under that. Tap Download and MAME auto installs on your Pandora. The search function will constantly be tweaked like that. We have a small enough software base to allow it.

Typing, say, 'ZX spectrum' might give the result 'There are two very good Spectrum emulators, these are : ... ', but typing 'PS2' might say 'The Pandora Cannot emulate or run PS2 software in any meaningful way, however someone did compile a PS2 emu, and if you really want to see it working at 1FPS here it is:...'.

This allows us to educate users quickly.

I've coded the system so that should the appstore become huge we can easily add areas and move software around.

[StreaK]

uh.. this is good, Craig. Love to see "+2 months from now" answer for question "when i'll get my pandora?"

btw..
AppStore could use keyboard-scripts like in http://www.keyboardr.com/
try to search something and start navigate using keyboard

Edited by StreaK, 30 December 2009 - 05:32 PM.

[craigix]

uh.. this is good, Craig. Love to see "+2 months from now" answer for question "when i'll get my pandora?"

That's a good idea, I might code in some answers to oddball questions.

[Trevsweb]

magic 8 ball type stuff lol

[emil10001]

Can you include tags in your setup as well? Maybe keep the categories, but augment that with the ability to filter on tags, or search on tags. That way, things that fit into several categories can be easily found.

[gadgetoid]

My horribly unpolished 2 cents based around the idea of tagging:

http://gadgetoid.com/pandora/store1.jpg
http://gadgetoid.com/pandora/store2.jpg

[second exodous]

I never thought of the Pandora not having many games, I guess that would affect how the store is search-able. I hope it will have more games, not just because it's more powerful but because it's more of a generic Linux and easier to port to.

[StreaK]

Ah, forgot to ask. Will appstore have 2 version of pages [ pandora specific 800x480 ] and [ other ] or only one dynamic resized?

[cobaltage]

I've been wondering lately if the app store can eventually take the form of a video game style interface controlled by the game controllers. Not like a video game itself, but like a loadout or options screen within a video game. Either built within a browser or maybe with a local frontend application that communicates with a server.... I just think that would be cool, and consistent with the Pandora's design.

I think its best solution for this. Maybe SDL + C++ CURL extensions could be good solution to do this [ completely independent frontend that dont require web browser running ]

I've also been wondering if some new elements in HTML5 will obviate the need for a special client frontend.

But, I guess the topic is pre-mature; the appstore definitely needs a web interface no matter what else happens, and it will have to have a definite structure, which Craig et al. are working out now; and HTML5 isn't out yet anyway.

Edited by cobaltage, 30 December 2009 - 06:25 PM.

[emil10001]

My horribly unpolished 2 cents based around the idea of tagging:

http://gadgetoid.com...dora/store1.jpg
http://gadgetoid.com...dora/store2.jpg

As far as I can tell, there are no tags in there. Just categories. My point was that when you break things down into categories, there will likely be certain things that don't fit neatly into just one category. Therefore, you can augment the categorization with tags, where multiple tag can be applied to each app. This allows the tags to get more specific than categories could be, while still allowing there to be categories for presentation purposes as well.

Example:
Title: Quake II
Category: Game
Tags: Game, FPS, Multiplayer

Title: Lerp
Category: Game
Tags: Game, Platformer, Puzzle, Exclusive

[B-ZaR]

I'm just throwing this out here. It's probably a dumb idea, but I want to share it anyways . The typical way to get around the similar tag problem (sidescroller, side-scroller...) is to show the user tags similar to those he's trying to use (if available). Because there usually are gazillions of tags, it's not really feasible to show the user all the tags to select from. However, some tag types, like those describing genre, actually form a subset of the tags that probably is of a manageable size.

This got me thinking about categorizing the tags into genre tags (shmup, platformer, fps), content tags (multiplayer, campaign, casual), context tags (scifi, zombies) and so on. Now I'm not suggesting having both categories and tags, but rather a way to organize the tags to aid in finding the correct pre-existing ones if there are any. New tags can of course be added like in any other system if needed. The submission system could show the user all the existing genre tags (because there are few), but use the search-as-you-type functionality for context tags (because there are many).

It could also be worth it to consider having the possibility to link tags to each other based on for example similarity (possibly synonymity) or same context (eg. elves and orcs). This way the user could find out tags and get suggestions on search words.

Just putting these here out in the open

EDIT: Oh, and one tag category could be type (game, application, utility...). See the post above.

Edited by B-ZaR, 30 December 2009 - 07:12 PM.

[gadgetoid]

My horribly unpolished 2 cents based around the idea of tagging:

http://gadgetoid.com...dora/store1.jpg
http://gadgetoid.com...dora/store2.jpg

As far as I can tell, there are no tags in there. Just categories.

Confirmed: Worst horizontal tag cloud ever.

[Laurencevde]

Why are we trying to reinvent the wheel if there are already perfectly well functioning systems out there that we can use, such as opendesktop.org and its accompanying sites, like kde-look.org, kde-apps.org, gnome-look, maemo-apps, and more.
Using that'll give us a fully functional app-store with all of the features we could wish with practically no development time...

[emil10001]

Why are we trying to reinvent the wheel if there are already perfectly well functioning systems out there that we can use, such as opendesktop.org and its accompanying sites, like kde-look.org, kde-apps.org, gnome-look, maemo-apps, and more.
Using that'll give us a fully functional app-store with all of the features we could wish with practically no development time...

I think that there are several reasons why it would be better to have a specifically developed app-store for the Pandora. First off, none of the options that you listed have anything to do with selling apps, at all. Second, we are going to be using a handheld device; such devices warrant an interface that minimizes the number of steps it takes to find the content that you are looking for. Third, I don't know if you have ever tried to find something specific on one of those sites, with only a vague idea of the name/category that you had found it in before, but it isn't always easy to find things. I remember looking for several wallpapers or window themes and spending more than 20 minutes searching because of the categorization of things on those sites. While I generally like the content on those sites, finding things is not always easy, partly due to the fact that they use categories and not tags, which is why I have suggested the addition of tagging to apps in the pandora app store.

[craigix]

Why are we trying to reinvent the wheel if there are already perfectly well functioning systems out there that we can use, such as opendesktop.org and its accompanying sites, like kde-look.org, kde-apps.org, gnome-look, maemo-apps, and more.
Using that'll give us a fully functional app-store with all of the features we could wish with practically no development time...

No it won't. I'm not being nasty saying that. Now I'm going to sound like DaveC:

We have a portable system which is unique and a scene which shapes what we want from an appstore.

That kde apps/maemo site is a joke. It's a total mess. Try browsing that at 800*480. It's unreadable and has one of the worst used/wasted space ratios I've ever seen. Half the screen is taken up by that empty top banner. Nothing a Pandora user wants to know is there. The search function is a tiny, clunky mess.

I already looked at free file archives. Anything we tried to do with them would just look like a hack. They are not designed to run on portable devices and definitely not designed with something like the Pandora scene in mind.

Wait until you see the Pandora appstore. You will understand why it had to be built from the ground up. You will see the advantages - and if you don't like something about it we can all fix it quickly. The chances are you will even be able to skin it.

We are doing internal testing on the appstore now and will be doing something public soon.

Suggestions are still welcome, I rewrote a huge chunk of it today based on a recent suggestion.