Sandboxed file access [LWN.net]

Sandboxed file access

Posted Sep 5, 2012 1:26 UTC (Wed) by khim (subscriber, #9252)
In reply to: Sandboxed file access by geofft
Parent article: Ubuntu's new app developer upload process proposal

I'm skeptical that app developers will be able to adapt to this new API and restrictions, and find it easier than just getting their app packaged the "normal" route (through Debian), which is what this proposal is attempting to sidestep.

"Usual route" is not acceptable for the wast majority of the applications. I mean commercial closed-source ones. Don't be misled by the stats which show that, e.g., most Android applications are "free": yes, they are free-as-in-beer, but that's because they include ads and these don't really fly in open-source world.

But yes, if there will be money to be had then applications developers will jump through necessary hoops. I'm skeptical too, but for a different reason: Ubuntu attracts a lot of "every app must be free!" guys thus it's not clear if there are enough money on the table to make it an attractive target for the app developers.

Other alternative (webapps) are more interesting: they make it possible to create "kind-of-native" applications by adding just a few tweaks to the existing web apps. This may fly: sure, the advantages are small, but the efforts required are small, too.

For the native apps... well, the effort to recompile everything under Linux is already huge! We've found about this the hard way: initially most NaCl developers were Linux guys, but as time goes on more and more come from MacOS and Windows world. They demand Visual Studio integration, etc. Native Client is cross-paltform so it works for us (not 100% well: core is cross-plaform, but WebGL is not so we frequently see problem with applications which were never even tested on Linux), but for Ubuntu this will be big problem: I doubt Canonical will rush to develop Visual Studio plugin any time soon.

(Log in to post comments)

Sandboxed file access

Posted Sep 5, 2012 1:33 UTC (Wed) by geofft (subscriber, #59789) [Link]

My understanding of this initiative is that it's specifically targetting free-software applications. The spec claims that the usual app review process for commercial software (whether gratis or not) is working fine, but the app review process for free software isn't scaling.

"... if the developer is selling their software, they are reasonably well catered: the consumer apps team will take care of them delivering their apps into Ubuntu. For Free Software developers ... they must do more work themselves. The Application Review Board (ARB) are there to review these apps but the ARB, as a team of volunteers, is not able to scale with increases in the queue ...."

In particular, this seems to mean that commercial software will be manually reviewed, not subject to this sort of sandboxing, and also not subject to the Debian / Debian-packaging route, which seems to work well for everyone. I'm not objecting to that.

Sandboxed file access

Posted Sep 6, 2012 9:59 UTC (Thu) by njwhite (subscriber, #51848) [Link]

There's also the issue that many people in our community find ads somewhat offensive, so would release a build of the program without them, which may well be more popular.

Sandboxed file access

Posted Sep 6, 2012 10:36 UTC (Thu) by krake (subscriber, #55996) [Link]

Well, that would mean somebody else is volunteering to do your packaging :)

Or one could just allow to switch it off. Users who'd like to support you would keep it on, user who don't or prefer direct donations could just treat it like adblock for browsers.

Anyway, the question was if there are technical reasons prohibiting ads supported Free Software

Sandboxed file access

Posted Sep 6, 2012 10:42 UTC (Thu) by nix (subscriber, #2304) [Link]

Certainly many of us find ads offensive *when badly done*. But I'm about as strongly against ads as you can find -- I block them almost everywhere because I have visual problems which mean that any ads displayed next to text will visually overlap the text, so if they're not of very similar colour and 'ink' density (which they never are) they'll make the text unreadable. And ads that leap out at you are unacceptable because they're a distraction and I have enough problems with *that* as it is.

But even for me *some* ads are tolerable. Google's text ads are of the same density as the text and are clearly delineated so I can skip them. Ads on the special-offer Kindle are acceptable because they're only displayed when the device is off and thus not displaying text I want to read anyway. And because the latter ads, at least, are actually targetted to some degree, sometimes those ads *are* useful, in that they tell me about things I would have wanted if I'd known they existed. If you'd told me three years ago that I'd avoid paying to turn off ads because the ads were useful I wouldn't have believed you. But no, it *is* possible to do advertising right enough that even someone with biological reasons to loathe advertising can like them. It's just that most online-ad people are so far from that that it is hard to believe.

Sandboxed file access

Posted Sep 6, 2012 11:37 UTC (Thu) by njwhite (subscriber, #51848) [Link]

> But I'm about as strongly against ads as you can find
> <snip>
> But even for me *some* ads are tolerable.

You're not as strongly against ads as me, then ;)

I'd argue adverts are never ever truly useful as recommendation agents. They are inherantly driven by who has the money to run them. So they work against small, new, innovative players. Let's not forget, our contemporary problem is *not* that we have too little information about new, interesting things. Advertising serves to shift the balance of the things we hear most about to those of entrenched interests.

Then there's the whole not wanting to be manipulated by said interests any more than you can avoid (and people tend to underestimate the degree to which marketing works, even on them.) Or surveilled by them.

Any business model that relies on ads is one I disapprove of. It pushes the initial economic cost to the reader down to zero, but at too great a cost.