▲ 19 points
back
17 comments
I'll call it first: BS.
Having worked in High Performance Computing, I can say that there is nothing worse than a heterogenous workload. If you've got image processing, data mining, number crunching, and simulations all running on the same hardware and OS, you miss out on a lot of optimization. If you are only doing one of those, then you can get quite a bit more performance out of the hardware by tuning it properly. SaaS, by it's specialized nature, allows that tuning to take place. Running everything on your own machines just slows them all down.
Though likely accurate, your rebuttal doesn't address Stallman's argument. Stallman doesn't claim that SaaS is cost-inefficient, slow, or impractical; he only claims that it should be avoided since it is a forfeiture of software/computation freedom.
While I wouldn't go as far to call it BS.
I agree with your argument.
I believe we need better laws that will ensure the freedom rather than modifying the system. In all other aspects of human life, we have relied on legal system to achieve economies of scale, while preserving our freedom. We dont grow our own food, or manufacture our own drugs, rather we rely on oversight of FDA etc. Sure the system isn't perfect but still its better than having a farm in your backyard and a chemical plant in your basement.
We need a legal revolution to address issues posed by rapidly evolving technology.
a.) Yes, there are some privacy and ownership/property problems that legislation could possibly solve, but…
b.) more laws makes it harder and less fun to write, distribute, and profit from making software.
However, I do think that a legal recourse for users to be able to export all of their data from a service in some format makes sense, especially when said data are entrusted to software startups that have a proclivity for being bought, sold, and swiftly unsupported or end-of-life'd at the drop of a hat.
Or, simpler, and less encumbering: how about a developer-initiated consortium of companies who "won't screw you over" when it comes time to back-up or find a new home for your data? Something like a Better (Software) Business Bureau who promises not to leave you high and dry… but now I'm just having excessively utopian thoughts.
I don't think the practical risk here justifies any such thing, because all of this is occurring in a market with possibly one of the lowest barriers-to-entry of any market in human history. With the ubiquity of screen scrapers, the power of various programming languages and environments, and the interconnectedness of the web, any attempt to jail your data can usually be met, quickly, by a large number of people.
For example: back when Facebook was really irritating everyone, Diaspora came along. Now, Diaspora should probably be considered a failed project at this point, but along the way there was a not-small number of developers that came forward and said, "hey, I already did this, start with my code...", and if the anger over Facebook hadn't subsided, I have no doubt that at least one of those projects would have become successful.
Or, for another example: when word got leaked that Yahoo was going to "sunset" Delicious, it took me all of a few bucks and a couple of clicks to export every last one of my bookmarks to Pinboard. (Thanks Pinboard!)
Proprietary software, by contrast, carried much greater risk, because it often used secret data formats which were extremely difficult (or impossible) to reverse-engineer, and because it was much more difficult to scrape the data from the screen in a reliable way. I have direct experience with this: small grocery stores tend to use a product called NCR ScanMaster, which at least semi-recently was "powered" by a truly atrocious Btrieve database backend, which was unreadable without the schema files, which nobody would give you without lots of money. It was damned hard to reverse-engineer too; I have a proud history of that kind of thing, and it totally stumped me for all but the simplest of data extractions.
One of my clients needed to be able to make special sale-related changes to large numbers of products on a regular basis without paying someone to sit there and key it in all day (and make mistakes while they were at it). We eventually came up with using AutoHotKey to control the screen and data fields, but even that was rather challenging because of ScanMaster's variable time delays between screens and AutoHotKey's limitations.
Contrast that with anything web-related, where I could use PHP or Javascript & Greasemonkey or one of any of a number of solutions to parse and scrape the content of any site within a matter of hours.
I can't read the article right now -- the site is down -- but the thing about Stallman is that he's a radical, and he will see his epic foe everywhere he looks. Cloud computing and SaaS have become popular, and are replacing many aspects of traditional desktop computing, but if he's making the claim that there is just as much danger, in terms of personal freedom, with cloud computing and SaaS as there was with proprietary desktop software or timesharing systems, well, he's just plain wrong.
He's also, in this context, probably completely irrelevant. I suspect that his opinions of cloud computing or SaaS will have epsilon impact on the industry, no matter how loudly or oft-stated.
I couldn't agree more.
Once you're scaling the best thing to do is build purpose built machines tuned to your workload. If you've got a bunch of ruby's running and your RAM is at 100% but CPU at 50% you can over spec your web front ends on RAM to fully utilize your CPU.
If you've got a DB machine, then you'll want to run 200-500 spindles with another 20 SSDs for indexes per socket and as much RAM as will fit.
Every day Stallman seems a little less unreasonable to me.
One reason you should not use web applications to do your computing is that you lose control. It's just as bad as using a proprietary program. Do your own computing on your own computer with your copy of a freedom-respecting program. If you use a proprietary program or somebody else's web server, you're defenceless. You're putty in the hands of whoever developed that software.
In the shadow of Facebook, who would disagree with this?
I am just not obsessed with total control. There are things I care about, and there are things I don't.
If you use a proprietary program or somebody else's web
server, you're defenceless
I disagree that I should care about who's computer I do my computing on. As long as I retain control of my data and the software, why should I care who is the owner of the metal box full of electronics that that software happens to be running on?
According to the article, social networking is not SaaS.
Stallman's argument assumes there is an equivalent non-SaaS version of every service or that you should live without it if you can't get it as open source. I am giving up absolute control to use Google's search engine but it is not practical to build my own system to index the entire web with sub-second response times. If there was a distributed open source search engine project where a large number of people could pool resources to recreate something similar to Google that would be a different story.
my god, you may have inadvertently had a great idea. There's a lot of cool things that could open up. For instance, i believe right now it is not kosher to scrape google results. However an open source distributed engine would be totally kosher.
Search results could potentially be more relevant... there would be millions of computers instead of hundreds of thousands. people could play with ranking algorithms... this is a cool thought.
I thought it was a cool idea too. Lack of control means manipulation of rankings might be a problem to solve. You would need to ensure that spammers couldn't flood the system with fake information.
Maybe its possible to use that to the networks advantage? If manipulating the data means you need to provide additional computation power/storage... maybe its possible to find a balance?
I would be interesting to use multiple random parties to check each other. For example, have 3 independent nodes index the same site and verify that no one has tampered with the data.
The link is dead for me, but googling what others quoted yields this Guardian article from 2008:
http://www.guardian.co.uk/technology/2008/sep/29/cloud.compu...
Markets for targeted SaaS solutions exists. Nobody is going to develop these for free and the businesses who need them are more than willing to pay (the value gained is much, much higher than the subscription fee). There are enough different businesses with the same need, that economically combined (through fees) they can support a team that develops the necessary software. These businesses are not going to collaborate with their competitors to develop the same software in a million years, but through SaaS providers, they basically do. Most of the people I know who work in software are basically involved in this at some level. It's not going away, and in many cases, it is really the only reasonable model.