Mike Linksvayer

Doug Cutting on Cloud: commodity or proprietary?:

As we shift applications to the cloud, do we want our code to remain vendor-neutral? Or would we rather work in silos, where some folks build things to run in the Google cloud, some for the Amazon cloud, and others for the Microsoft cloud? Once an application becomes sufficiently complex, moving it from one cloud to another becomes difficult, placing folks at the mercy of their cloud provider.

I think most would prefer not to be locked-in, that cloud providers instead sold commodity services. But how can we ensure that?

If we develop standard, non-proprietary cloud APIs with open-source implementations, then cloud providers can deploy these and compete on price, availability, performance, etc., giving developers usable alternatives.

That’s exactly right. Cloud providers (selling virtualized cpu and storage) are analogous to hardware vendors. We’re in the pre-PC era, when a developer must write to a proprietary platform, and if one wants to switch vendors, one must port the application.

But such APIs won’t be developed by the cloud providers. They have every incentive to develop proprietary APIs in order to lock folks into their services. Good open-source implementations will only come about if the community makes them a priority and builds them.

I think this is a little too pessimistic. Early leaders may have plenty of incentive to create lockin, but commoditization is another viable business model, one that could even be driven by a heretofore leading proprietary vendor, e.g., the IBM PC, or Microsoft-Yahoo!

Of course the community should care and build the necessary infrastructure so that it is available to enable a potential large cloud provider to pursue the commoditization route and to provide an alternative so long as no such entity steps forward.

Cutting has been working on key parts of the necessary infrastructure; read the rest of his post for more.

I strongly prefer voluntary action. However, software patents are not amenable to workaround and so must be attacked directly through less savory legal, legislative, and electoral routes (though if software patents are toxic to free software, the opposite is also true, so simply creating and using free software is a voluntary if indirect attack on software patents).

Software patents are the major reason multimedia on the web (and on computers generally) is so messed up — few multimedia formats may be implemented without obtaining many patent licenses, and amazingly, this is sometimes impossible:

[The MPEG-4 framework] is so patent-encumbered that today no one really knows who has “rights” to it. Indeed, right now, no new MPEG-4 licenses are even being issued.

As the End Software Patents site emphasizes, software patents negatively impact every sector now that everything uses software.

My only problem with the ESP site (and many others, this is just a general peeve of mine) is that it does not even link to similar resources with a non-U.S. jurisdiction focus. For example, the What Can I Do? page might state that if one is reading the page but not in the U.S. (because that never happens), please check out FFII (EU) and similar.

In any case, please join the effort of ESP and others to eradicate software patentsweapons of mass destruction. Ars Technica has a good introductory article on ESP.

I continue to be intrigued by LimeWire’s huge and relatively unsung popularity. According to a December 13 release:

More than one-third of all PCs worldwide now have LimeWire installed, according to data jointly released by Digital Music News and media tracking specialist BigChampagne. The discovery is part of a steady ascent for LimeWire, easily the front-running P2P application and the target of a multi-year Recording Industry Association of America (RIAA) lawsuit. For the third quarter of this year, LimeWire was found on 36.4% of all PCs, a figure gleaned from a global canvass of roughly 1.66 million desktops.

The installation share is impressive, and unrivaled. But growth has actually been modest over the past year. LimeWire enjoyed a penetration level of 34.1% at the same point last year, a difference of merely 2.3%.

These figures don’t jibe with those supposedly from the same parties from earlier this year, which found LimeWire installed on 18.63% of desktops. A writer on TorrentFreak who has presumably seen the more recent report (US$295, apparently including the requisite section titled “LimeWire Challenged by…Google?”) says:

From the data where the report is based on we further learn that Limewire’s popularity is slowly declining. However, with an install base of almost 18% it is still the P2P application that is installed on most desktop computers. Unfortunately Digital Music News has trouble interpreting their own data, they claim in their press release that it is 36.4%, but that is the market share compared to other P2P clients (shame on you!).

In other open source filesharing application news, Shareaza made its first release in over two years on December 1.

Via Slyck.

The incredibly productive Kragen Sitaker, in Exegesis of “Re: [FoRK] Calling [redacted] and all the ships at sea.”:

The major political issue of today [0] is that music distribution companies based on obsolete physical-media-distribution models (”record labels”) are trying to force owners of new distribution mechanisms, mostly built on the internet, to pay them for the privilege of competing with them; the musical group “The Grateful Dead” used to permit their fans to distribute their music by making copies of taped performances, and most of the money the Dead made came from these performances; it is traditional for performances not to send any revenue to the record label. Long compares the record labels to buggy-whip manufacturers, who are the standard historical symbol for companies who went out of business because of technological change.

This clearly relates to the passage the footnote is attached to, which is about the parallel between Adam Smith’s economic “invisible hand” and the somewhat more visible hand that wrote the king’s doom on the wall in Daniel; in this case, the invisible hand has written the doom of the record companies on the wall, and their tears will not wash out a word of it. What this has to do with Huckleberry Finn’s prohibition on seeking symbolism or morals in the book, I don’t know, although clearly Huckleberry Finn’s prohibition relates to mortals hiding messages in texts.

[0] Yes, this means I think this is more important than the struggle over energy, or the International Criminal Court, or global warming, or nuclear proliferation — the issue is whether people should be permitted to control the machines they use to communicate with one another, in short, whether private ownership of 21st-century printing presses should be permitted. (Sorry my politics intrude into this message, but I thought “the major political issue of today” required some justification, but needs to be there to explain the context to people reading this message who don’t know about it.)

That will probably seem a pretty incredible claim, but I often agree, and think Sitaker understates the case. Music distribution companies are only one of the forces for control and censorship. The long term issue is bigger than whether private ownership of 21st-century printing presses should be permitted. The issue is whether individuals of the later 21st-century will have self-ownership.

The Wikimedia Foundation board has passed a resolution that is a step toward Wikipedia migrating to the Creative Commons Attribution-ShareAlike license. I have an uninteresting interest in this due to working at Creative Commons (I do not represent them on this blog), but as someone who wants to see free knowledge “win” and achieve revolutionary impact, I declare this an important step forward. The current fragmentation of the universe of free content along the lines of legally incompatible but similar in spirit copyleft licenses delays and endangers the point at which that universe reaches critical mass — when any given project decides to use a copyleft license merely because then being able to include content from the free copyleft universe makes that decision make sense. This has worked fairly well in the software world with the GPL as the copyleft license.

Copyleft was and is a great hack, and useful in many cases. But practically it is a major barrier to collaboration in some contexts and politically it is still based on censorship. So I’m always extremely pleased by any expansion of the public domain. There could hardly be a more welcome expansion than Daniel J. Bernstein’s release of his code (most notably qmail) into the public domain. Most of the practical benefit (including his code in free software distributions) could have been achieved by released under any free software license, including the GPL. But politically, check out this two minute video of Bernstein pointing out some of the problems of copyright and announcing that his code is in the public domain.

Bernstein (usually referred to as ‘djb’) also recently doubled the reward for finding a security hole in qmail to US$1,000. I highly recommend his Some thoughts on security after ten years of qmail 1.0, also available as something approximating slides (also see an interesting discussion of the paper on cap-talk).

I imagine thousands of bloggers have commented on gOS, a Linux distribution featuring shortcuts to Google web application on the desktop and preloaded on a PC sold (out) for $200 at Wal-Mart. Someone asked me to blog about it and I do find plenty interesting about it, so thus this post.

I started predicting that Linux would take over the desktop in 1994 and stopped predicting that a couple years later. The increasing dominance of web-based applications may have me making that prediction again in a couple more years, and gOS is a harbinger of that. Obviously web apps make users care less about what desktop operating system they’re running — the web browser is the desktop platform of interest, not the OS.

gOS also points to a new and better (safer) take on a PC industry business model — payment for placement of shortcuts to web applications on the desktop (as opposed to preloading a PC with crapware) — although as far as I know Google isn’t currently paying anything to the gOS developers or Everex, which makes the aforementioned cheap PC.

This is highly analogous to the Mozilla business model with a significant difference: distribution is controlled largely by hardware distributors, not the Mozilla/Firefox site, and one would expect end distributors to be the ones in a position to make deals with web application companies. However, this difference could become muted if lots of hardware vendors started shipping Firefox. This model will make the relationship of hardware vendors to software development, and particularly open source, very interesting over the next years.

One irony (long recognized by many) is that while web applications pose a threat to user freedoms gained through desktop free and open source software, they’ve also greatly lowered the barriers to desktop adoption.

By the way, the most interesting recent development in web application technology: Caja, or Capability Javascript.

Peter McCluskey:

If an AI started running in 2003 that has accumulated the knowledge of a 4-year old human and has the ability to continue learning at human or faster speeds, would we have noticed? Or would the reports we see about it sound too much like the reports of failed AIs for us to pay attention?

How old would a human child have to be to detect current spam nearly flawlessly (given lots of training)? To write spam that really does seem to be from your kids?

If Gmail accounts essentially stop getting spam, there’s a child AI is at Google^*. If spam stops being largely non- or pseudo-sensical, a child AI lives in a botnet.

^*Most likely AI host, or so some outside the Singularity Summit seemed to think; previous post on that event.

Steve Omohundro’s talk at today’s Singularity Summit made the case that a self-improving machine would be a rational economic actor, seeking to eliminate biases that get in the way of maximizing its utility function. Omohundro threw in one purely speculative method of self-preservation — “energy encryption” — by which he meant that an entity’s energy would be “encrypted” such that it could not be used by another entity that attacks in order to get access to more energy.

I note “energy encryption” here because it sounds neat but seems impossible and I can find no evidence of use in this way before Omohundro (there is a crypto library with the name).

The “seems impossible” part perhaps means the concept should not be mentioned again outside a science fantasy context, but I realized the concept could perhaps be used with artistic license to describe something that has evolved in a number of animals — prey that is poisonous, or tastes really bad. What’s the equivalent for the hypothetical Jupiter Brain in a dangerous part of the galaxy? A stock of antimatter?

I also found one of Omohundro’s other self-preservation strategies slightly funny in the context of this summit — a self-aware AI will (not should, but as a consequence of being a rational actor) protect its utility function (”duplicate it, replicate it, lock it in safe place”), for if the utility function changes, its actions make no sense. So, I guess the “most important question facing humanity” is taken care of. The question, posed by the Singularity Institute for Artificial Intelligence, organizer of the conference:

How can one make an AI system that modifies and improves itself, yet does not lose track of the top-level goals with which it was originally supplied?

I suppose Omohundro did not intend this as a dig at his hosts (he is an advisor to SIAI) and that my interpretation is facile at best.

Addendum: Today Eliezer Yudkowsky said something like Omohundro is probably right about goal preservation, but current decision theory doesn’t work well with self-improving agents, and it is essentially Yudkowsky’s (SIAI) research program to develop a “reflective decision theory” such that one can prove that goals will be preserved. (This is my poor paraphrasing. He didn’t say the words “reflective decision theory”, but see hints in a description of SIAI research and a SL4 message.)

There’s been a fair bit written about ‘Moore’s law for software’, usually complaining that there isn’t one. My guess is that’s nuts, but I’d love to see some rigorous analysis (I bet I’m just ignorant of it).

Interesting tidbit from San Jose Mercury-News article two weeks ago Penny-pinching entrepreneurs changing world of venture capital:

Ten years ago, six or seven programmers would have been needed to achieve the results of one programmer today, valley veterans say.

If true, that’s an annual increase in programmer productivity of about twenty percent. Let’s say it’s actually half that due to exaggeration or Brooks’ Law (adding headcount to a software project doesn’t scale well–though on second thought Brooks’ Law could magnify productivity increases, by allowing teams to get smaller). That would make for a doubling time of about seven years. Not nearly as impressive as Moore’s Law doubling of transistor density every two years, but still exponential. And my wild guess is that it has been fairly consistent over the history of programming.

For my five year old impressions on the matter, see this thread.

Addendum: Depending (in part) on how far back you consider the history of programming to go, of course a consistent doubling time for software (or hardware) doesn’t make sense, but rather accelerating change. Doubtless Ray Kurzweil has many graphs attempting to demonstrate this for software in his books. I didn’t intend to go there in this post, but it is timely, as I’ll probably attend the Singularity Summit in a couple weekends.

LimeWire is supposedly installed on nearly one in five PCs. “Current installation share” for filesharing programs according to BigChampagne and PC Pitstop:

1. LimeWire (18.63%)
2. Azureus (3.43%)
3. uTorrent (3.07%)
4. BitTorrent (2.58%)
5. Opera (2.15%)
6. Ares (2.15%)
7. BitComet (1.99%)
8. eMule (1.98% )
9. BearShare (1.64%)
10. BitLord (1.38%)

It’s a little odd to include all those BitTorrent clients, given their very different nature. All but LimeWire, Ares, eMule, and BearShare are BT-only (their P2P download component — Opera is mainly a web browser, with built in BT support). Recent versions of LimeWire and Ares also support BT, so another provocative headline would be “LimeWire the most popular BitTorrent client?”

Most recent (for surveys publishing numbers in 2007) usage share for Firefox ranges from 11.69% to 14.32%. Of course usage share is very different from installation share (compare Opera installation share above at 2.15% and recent usage share between 0.58% and 0.77%) and P2P filesharing and download clients have different usage patterns, so any comparison is apples to oranges. However, if one could extrapolate from the Opera numbers for installation and usage, LimeWire is not more popular than Firefox.

LimeWire is still impressively popular. This probably is mostly due to open source being less susceptible to censorship than proprietary software (which has a half-life shortened by legal attack in the case of P2P). Still, I’d like to see LimeWire gain more recognition as an open source success story than it typically gets.

The really interesting speculation concerns how computing (and ok, what may or may not have been called Web 2.0) would have been different had P2P not been under legal threat for seven or so years. Subject for another post. We can’t go back, but I think it’s very much worth trying to get to a different version of there.

Yes, I know about significant digits. I’m just repeating what the surveys say.

Philip Greenspun reporting from the Digital Freedom Exposition in South Africa:

My personal view is that it is not the job of computer nerds to keep people free of disease. We build interesting Web sites and other services to make life interesting and worth living as long as the biologists and doctors are able to keep folks alive. Even if human life expectancy were reduced to 30 years, we shouldn’t abandon our keyboards and move into the medical labs since even a 30-year life can be significantly enriched with Google and Wikipedia.

He understands comparative advantage.

Another fun excerpt from the same post:

More than food, shelter, or other seeming essentials, they wanted Internet access, starting with an Internet cafe for women in the capital (under the Islamic regime, only men were allowed to visit Internet cafes).

Double whammy on those who complain that others who attempt to bring technology to the poor should focus on basic needs first.

I initially installed Ubuntu Linux 5.10 on my new Dell Inspiron 6000 in November, 2005. I fully expect it to begin having assorted hardware problems this year with the amount of use I give it, though hopefully using an external keyboard (excellent, and only $4.99) and mouse at work will extend its life.

I upgraded to 6.06 shortly after its release but didn’t get around to blogging it. I encountered two problems:

• I had to re-install i915resolution to get back full 1920×1200 screen resolution (1600×1200 without it)
• Resume from suspend-to-memory was broken. A fix in the form of an upgraded acpi-support package was available in the next couple days.

I didn’t upgrade to 6.10 upon release, mostly because I just didn’t get around to it. I saw a couple days ago that 7.04beta is out, so I finally upgraded to 6.10. The upgrade process went without a hitch, but resume from suspend-to-memory was broken again. I’m sure there’s an easy fix, but I decided to take the plunge and upgrade to 7.04beta.

Everything now works, though as Brad Templeton notes the upgrade process stopped several times, waiting on my answers to unnecessary questions. Apart from this annoyance, which hopefully will be fixed before 7.04 gets out of beta, I remain impressed with Ubuntu and how it is progressing. I’m also reminded of how rapidly free software projects are doing new feature releases. I don’t think there are any visible applications I use on this laptop that haven’t been significantly improved in the last 17 months.

Ed Felten asks How Much Bandwidth is Enough? (emphasis added):

It is a matter of faith among infotech experts that (1) the supply of computing and communications will increase rapidly according to Moore’s Law, and (2) the demand for that capacity will grow roughly as fast. This mutual escalation of supply and demand causes the rapid change we see in the industry.

Funny how that seems to happen.

Thus far, whenever more capacity comes along, new applications are invented (or made practical) to use it. But will this go on forever, or is there a point of diminishing returns where more capacity doesn’t improve the user’s happiness?

There’s always a point at which purchasing more bandwidth doesn’t make sense given the price of bandwidth and other goods. But will there ever be a point at which more bandwidth, even at zero price, has no utility? I doubt it.

There is a plausible argument that a limit exists. The human sensory system has limited (though very high) bandwidth, so it doesn’t make sense to direct more than a certain number of bits per second at the user. At some point, your 3-D immersive stereo video has such high resolution that nobody will notice any improvement. The other senses have similar limits, so at some point you have enough bandwidth to saturate the senses of everybody in the home. You might want to send information to devices in the home; but how far can that grow?

Human sensory system? The home? By the time there is enough bandwidth to max out the human sensory system and auxiliary devices humans will not be important on the scene.

Threats and Mitigation appeared on the OLPC wiki a couple weeks ago. It closes with very brief calls for capability security and agoric computing, unsurpsingly, considering the source.

But I wanted to point out the article’s proposal for mitigating social engineering:

The best place to defeat the hoax is in the mind of the intended victim. How? With educational tools shipped on the OLPC itself. Suppose the computer had a training course that taught each student-owner how to run the hoax himself.

This strikes a chord with me because I already think “we” (artists, bloggers, programmers, preachers, friends — see friends don’t let friends click spam) should promote not engaging spammers and scammers and because I’m annoyed by the practice of computer vendors (HP/Compaq anyway) pre-loading consumer Windows machines with scads of “special offer” programs that are annoyances at best and would fairly be considered malware if they didn’t come preinstalled.

Instead of bombarding a new user with vendor-approved spam the first time a computer is turned on an enlightened consumer PC vendor (I include OLPC here) would show a brief safe computing video. Support costs may even be reduced through such a move.

On the technical side OLPC posted a summary of their Bitfrost security platform. While much is left to the imagination at this point (there’s an annoying lack of references or even buzzwords in the specification), it sounds like OLPC programs could get a whole lot less authority than those on any mass platform so far.

Tom Evslin says Apple Fails to Reinvent Telecommunications Industry:

Steve Jobs claims that iPhone will “reinvent” the telecommunications sector. Wish it were so but it ain’t!

The design of the phone – no hard buttons, all touch on screen, sounds like everything we expect from Steve and from Apple: it’s all about the GUI and that part’ll be fun. But the business relationship is as old school as it can get: exclusive US distributorship through Cingular
…
Short term this is a good tactic for Apple because it protects the iPod franchise for a while. Long term I think it’s terrible strategy. It invites an endrun from someone who IS willing to reinvent the industry or simply allies themselves with a Cingular competitor.
…
Remember how wonderful the Mac GUI was? But it only ran on machines from Apple. Remember how crappy Windows was at first? But it ran on machines from everyone and their brother. And now there’s Linux – even less restricted – running on anything that moves. Tell me again why it makes sense to have a phone that runs only on a service from at&t (in the US).

Like other Apple products, the iPhone is eye candy (ugly to me), but not revolutionary.

It looks like the FIC Neo1973, showing at CES, due to ship this quarter for US$350, running the OpenMoko platform (presentation), will be more in the right direction — unlocked and open for developers. Andy on the openmoko list has a very early comparison.

The Neo1973’s big missing feature, at least initially, is apparently Wi-Fi, due to a lack of open drivers. As a late adopter of gadgets, I can wait. I acquired my first and only mobile phone in 2003, and it’s easy on the eyes.

That said, I’d really like Portable online by 2010 to be true:

This claim is judged YES if and only if, by January 1, 2010, in any state with more than 5 million inhabitants, at least 25% of the adult population are “portably online”. A “state” can be a country or a member state in a federation.

Read more for how “portably online” is defined (the contract was written in 1995). My current guess (and the market’s; last trade at 30) is that without a more significant smartphone revolution than we’re seeing the criteria won’t be met before 2010, but not terribly long after.

OpenMoko via Jon Phillips. Second word in post title refers to a silly slide found at Engadget.

Luis Villa:

I’m boggled that Fedora, OpenSuse, and Ubuntu, all of whom have open or semi-open build systems now, are not actively seeking out Emperor and companies like Emperor, and helping them ship distros that are as close to upstream- and hence most supportable- for everyone. Obviously it is in RH, Canonical, and Novell’s interests to actively pursue Big Enterprise Fish like HP and Dell. But I’m really surprised that the communities around these distros haven’t sought out the smaller, and potentially growing, companies that are offering computers with Linux pre-installed.

Sounds exactly right to me. I’ve been thinking something similar for awhile, but as the post title suggests, focused on hardware vendors. Tons of them compete to sell Linux servers at the very low and very high ends and everything inbetween, but if you want a pre-installed Linux laptop you need to pay a hefty premium for slightly out of date hardware from someone like Emperor Linux. It seems like there’s an opportunity for a hardware vendor to sell a line of Linux laptops that aren’t merely repurposed Windows machines. It has seemed like this for a something like a decade though, and as far as I know HP and a couple others have only tentatively and temporarily offered a few lame configurations.

So I’d like to see a hardware startup (or division of an existing company) sell a line of laptops designed for Linux, where everything “just works” just as it does on Macs, and for the same reasons — limited set of hardware to support, work on the software until it “just works” on that hardware. There’s probably even some opportunity for Apple-like proprietary control over some aspects of the hardware. Which reminds me, what legal barriers, if any, would someone who wants to manufacture the OLPC design face? There is discussion of a commercial subsidiary for the project:

The idea is that a commercial subsidiary could manufacture and sell a variation of the OLPC in the developed world. These units would be marked up so that there would be a significant profit which can be plowed into providing more units in countries who cannot afford the full cost of one million machines.

The discussions around this have talked about a retail price of 3× the cost price of the units.

In any case Villa is right, distributions should be jumping to support hardware vendors, both the mundane and innovative sorts. Which Red Hat/Fedora is doing in the case of OLPC.

Update 20060926: In comments below Villa points out system76, which approaches what I want, excpet that their prices are mediocre and they don’t offer high resolution displays, which I will never do without again. David points out olpcnews.com, which looks like reasonable independent reporting on OLPC. I asked on the OLPC wiki about other manufacturers’ use of the design.

Brief thoughts on last week’s LinuxWorld Conference and Expo San Francisco.

Lawrence Lessig’s opening keynote pleased the crowd and me. A few points fof interest:

• Free speech is a strong aspect of free culture and at least implicitly pushed for a liberal interpretation of fair use, saying that the ability to understand, reintepret and remake video and other multimedia is “the new literacy” and important to the flourishing of democracy.
• The “read/write Internet”, if allowed to flourish, is a much bigger market than the “read only Internet.”
• Support free standards and free software for media, including Ogg and Gnash.
• In 1995 only crazies thought it possible to build a viable free software operating system (exaggeration from this writer’s perspective), now only crazies think wireless can solve the last mile competition problem. Go build free wireless networks and prove the telcos and pro-regulation lawyers (including the speaker) wrong.
• One of the silly video mashups Lessig played was Jesus Will Survive, featuring an adult Jesus in diapers hit by a bus. A few people left the auditorium at this point.

I’ve at least visited the exhibition space of almost every LWCE SF (the first one, actually in San Jose, was the most fun — Linus was a rock star and revolution was in the air) seemed bigger and more diverse, with most vendors pushing business “solutions” as opposed to hardware.

By far the most interesting exhibition booth to me was Cleversafe, an open source dispersed storage project that announced a Linux filesystem interface at the conference and was written up in today’s New York Times and Slashdot. I’ve been waiting for something like this for a long time, particularly since Allmydata is not open source and does not support Linux.

Also, Creative Commons won a silly “Best Open Source Solution” show award.

Addendum 20080422: If you’re arriving from an unhinged RedState blog post, see Lessig’s response.

Luis Villa on my constitutionally open services post:

It needs a catchier name, but his thinking is dead on- we almost definitely need a server/service-oriented list of freedoms which complement and extend the traditional FSF Four Freedoms and help us think more clearly about what services are and aren’t good to use.

I wasn’t attempting to invent a name, but Villa is right about my aim — I decided to not mention the four freedoms because I felt my thinking too muddled to dignified with such a mention.

Kragen Sitaker doesn’t bother with catchy names in his just posted draft essay The equivalent of free software for online services. I highly recommend reading the entire essay, which is as incisive as it is historically informed, but I’ve pulled out the problem:

So far, all this echoes the “open standards” and “open formats” discussion from the days when we had to take proprietary software for granted. In those days, we spent enormous amounts of effort trying to make sure our software kept our data in well-documented formats that were supported by other programs, and choosing proprietary software that conformed to well-documented interfaces (POSIX, SQL, SMTP, whatever) rather than the proprietary software that worked best for our purposes.

Ultimately, it was a losing game, because of the inherent conflict of interest between software author and software user.

And the solution:

I think there is only one solution: build these services as decentralized free-software peer-to-peer applications, pieces of which run on the computers of each user. As long as there’s a single point of failure in the system somewhere outside your control, its owner is in a position to deny service to you; such systems are not trustworthy in the way that free software is.

This is what has excited about decentralized systems long before P2P filesharing.

Luis Villa also briefly mentioned P2P in relation to the services platforms of Amazon, eBay, Google, Microsoft and Yahoo!:

What is free software’s answer to that? Obviously the ’spend billions on centralized servers’ approach won’t work for us; we likely need something P2P and/or semantic-web based.

Wes Felter commented on the control of pointers to data:

I care not just about my data, but the names (URLs) by which my data is known. The only URLs that I control are those that live under a domain name that I control (for some loose value of control as defined by ICANN).

I hesitated to include this point because I hesitate to recommend that most people host services under a domain name they control. What is the half-life of http://blog.john.smith.name vs. http://johnsmith.blogspot.com or js@john.smith.name vs. johnsmith@gmail.com? Wouldn’t it suck to be John Smith if everything in his life pointed at john.smith.name and the domain was hijacked? I think Wes and I discussed exactly this outside CodeCon earlier this year. Certainly it is preferable for a service to allow hosting under one’s own domain (as Blogger and several others do), but I wish I felt a little more certain of the long-term survivability of my own [domain] names.

This post could be titled “freedom needs P2P” but for the heck of it I wanted to mirror “free culture needs free software.”

Luis Villa provokes, in a good way:

Someone who I respect a lot told me at GUADEC ‘open source is doomed’. He believed that the small-ish apps we tend to do pretty well will migrate to the web, increasing the capital costs of delivering good software and giving next-gen proprietary companies like Google even greater advantages than current-gen proprietary companies like MS.

Furthermore:

Seeing so many of us using proprietary software for some of our most treasured possessions (our pictures, in flickr) has bugged me deeply this week.

These things have long bugged me, too.

I think Villa has even understated the advantage of web applications — no mention of security — and overstated the advantage of desktop applications, which amounts to low latency, high bandwidth data transfer — let’s see, video, including video editing, is the hottest thing on the web. Low quality video, but still. The two things client applications still excel at are very high bandwidth, very low latency data input and output, such as rendering web pages as pixels. :)

There are many things that can be done to make client development and deployment easier, more secure, more web-like and client applications more collaboration-enabled. Fortunately they’ve all been tried before (e.g., Java Web Start, JXTA, XUL, others of varying relevance), so there’s much to learn from, yet the field is wide open. Somehow it seems I’d be remiss to not mention XMPP, so there it is. Web applications on the client are also a possibility, though typical web application platforms only address ease of development and not manageability at all.

The ascendancy of web applications does not make the desktop unimportant any more than GUIs made filesystems unimportant. Another layer has been added to the stack, but I am still very happy to see any move of lower layers in the direction of freedom.

My ideal application would be available locally and over the network (usually that means on the web), but I’ll prefer the latter if I have to choose, and I can’t think of many applications that don’t require this choice (fortunately email is one of them, or close enough).

So what can be done to make the web application dominated future open source in spirit, for lack of a better term?

First, web applications should be super easy to manage (install, upgrade, customize, secure, backup) so that running your own is a real option. Applications like WordPress and MediaWiki have made large strides, especially in the installation department, but still require a lot of work and knowledge to run effectively.

There are some applications that centralizaton makes tractable or at least easier and better, e.g., web scale search, social aggregation — which basically come down to high bandwidth, low latency data transfer. Various P2P technologies (much to learn from, field wide open) can help somewhat, but the pull of centralization is very strong.

In cases were one accepts a centralized web application, should one demand that application be somehow constitutionally open? Some possible criteria:

• All source code for the running service should be published under an open source license and developer source control available for public viewing.
• All private data available for on-demand export in standard formats.
• All collaboratively created data available under an open license (e.g., one from Creative Commons), again in standard formats.
• In some cases, I am not sure how rare, the final mission of the organization running the service should be to provide the service rather than to make a financial profit, i.e., beholden to users and volunteers, not investors and employees. Maybe. Would I be less sanguine about the long term prospects of Wikipedia if it were for-profit? I don’t know of evidence for or against this feeling.

Consider all of this ignorant speculation. Yes, I’m just angling for more freedom lunches.

Apple’s penetration of the geek market over the last five years or so has bugged me … for that long. It has been far longer than that since I’ve read a comp.*.advocacy threadflamewar, so stumbling upon Mark Pilgrim’s post on dumping Apple and its heated responses made me feel good and nostalgic.

Tim Bray (who does not b.s.) answers Time to Switch? affirmatively.

I hope this is the visible beginning of a trend and that in a few years most people who ought to know better will have replaced laptops sporting an annoying glowing corporate logo with ones sporting Ubuntu stickers.