Are today's online services consistent with GNOME's mission?

Posted Jul 26, 2007 10:01 UTC (Thu) by ortalo (subscriber, #4654) [Link]

I totally agree with you on this. (Not withstanding the fact that, personnally, I am not at all convinced that this online desktop metaphor would constitute a positive evolution of computing.)
I also think that dogmatic positions should be avoided at all cost, especially with respect to the centralized vs. distributed, or remote vs. local visions of software architecture. In fact, it is my humble opinion that those defending these dogms always carefully hide the fact that, for a given task, one choice may be better than another; but that for all possible tasks, the right choice certainly is not always the same.

Applying that pragmatism to the OD idea and your remark about free (as in freedom) web services I'd like to share another personal opinion.
I wonder if public services is not *the* area where such a free software solution for fully online services would really be worth the effort.
Centralized management of data by public administrations is nearly unavoidable as the advantages of (technical) decentralization are not usually worth the cost for such big database management systems. Similarly, even though opening access to the widest users base possible is certainly useful (the famous "e-administration"), it is not really possible to offer something else than online and remote (carefully controlled) applications for such procedures which end-to-end control, in fine, *must* remain in the hand of public institutions. (I am speaking of really basic public services, like citizens identification, etc.)

Of course, nowadays, these online services are usually offered using proprietary software (and that extends to the client software too)... But imagine a full free software solution allowing states and their citizens to manage some administration procedures and documents online on a remote server.
If such a free software solution was available in fields such as demography, vehicles, taxes, safety services, etc., public institutions would probably consider it.[2] Whether they like it or not. For strong ethical reasons, but also for that very pragmatic obligation [1] which is the other side of "free". ;-)

We do not need to re-implement all the software coded by governements during 3 decades before gaining momentum on this. A few carefully chosen cases could prove useful already. But we need to write them entirely.
It could prove boring certainly to implement a boring procedure in its entirety, but I am sure there can be some fun too for those with a passion for multi-terabytes databases or careful anonymizing cryptographic algorithms...

[1] In my mother language it reads: "le souci des deniers publics"
[2] It has already started in some countries, but if I am not mistaken, it were not really fully community-driven initiatives...

Posted Jul 26, 2007 14:23 UTC (Thu) by hp (subscriber, #5220) [Link]

There is a free server side involved here. I've been hacking on online.gnome.org all week (it's not up yet, but see mockups), and Luis Villa is spending the balance of his Red Hat internship tackling the Free Service Definition problem. We've also got people from Creative Commons and Software Freedom Law Center and others helping out. We'll use online.gnome.org to be the first ones to adopt this and once we define what a free server side is and have our first free server side functionality, we can go from there.

We are also going to insist that the free online.gnome.org is the only "required" part of Online Desktop - I say "required" in quotes, since Online Desktop itself is an optional mode of GNOME, not a required thing.

The GNOME project is mostly a client-side project and few people in it know how to code server side stuff, so I think GNOME will mostly not be about the server side for now. But of course I would love to see free software server apps and there are tons of open source developers who know how to code them. I hope they will. I hope GNOME developers will learn eventually, too.

The essence of the GNOME Online Desktop client side to me should be choice - ability to use any server apps. That will preserve our ability to do free server apps.

A number of people have talked about using P2P or federation or <insert hand waving here> to avoid the hosting issue. While I'd love to see that, the fact is that it's insanely difficult and something of a research problem. As a result, I think it becomes an excuse; "we can ignore server apps, because they suck, we should use P2P!" - I'll be in favor of that when I see running, production, Internet-scale P2P code working with the same user benefits users are getting from server apps. Until then I think P2P is more wishful thinking than anything else. (Don't get me wrong, it works in limited domains and for limited kinds of functionality, but it seems to be impractical to replicate many of the server apps people use via P2P.)

The way to think of this is that if we want open source to be popular and want users to have a completely open source experience, we have to offer the same benefits users are getting from these server apps today. No excuses.

There is absolutely no way that allowing our client OS to become increasingly irrelevant to how people are using computers is helpful to free software. If we have a client OS that's highly relevant to what people are doing, in contrast, and people thus use the client OS, then we have the ability in the future to (for example) set the default server apps to free server apps.

If Microsoft owns the client OS, then you can imagine what the default server apps will be.