Robert Pogson

Change happens. It’s inevitable. Ubuntu is not immune. Individuals are not wheels that turn forever in a machine. They grow and move on. It is time to fill the spot left by Jane Silber. Matt Asay has been chosen. Is he the right person for the job? Time will tell, but he has decent creds: Novell and Alfresco. He is well known from his blog, The Open Road. A big plus may be that he is known and knows the USA. This is important because the USA is different. While known as a technological innovator, the USA has hangups about M$. They like the home boy who does good and creates an empire out of nothing. They like that so much they tend to overlook matters of anti-competitive acts and insecure IT which M$ causes. They see M$ as a generator of wealth, not as an infinite sink of costs. The rest of the world sees things much differently and are far ahead of the USA in FLOSS adoption even though many of the core components of FLOSS began in the USA: GNU, GPL, and several important distros. The USA accepts widely held myths about FLOSS such as 1% market share for GNU/Linux, freeloading, boxes with GNU/Linux being converted to that other OS, immunity from malware is due to insufficient share, etc. Further, the USA largely ignores the explosive developments of IT in other parts of the world, eg. China. USA adoption of GNU/Linux in business has largely been RedHat/Suse. Can Matt Asay make a difference to perception of GNU/Linux in the USA and can he make a difference in Ubuntu in some way to improve acceptance in the USA?

I have some doubts. The USA is locked up very tightly as the stronghold of empire. M$ gets everything its own way from lobbying, astroturfing, ignoring anti-competitive acts for ten years or more, and billions of promotional dollars supplied by monopolistic prices. There is a long uphill road for any FLOSS business to crack the USA market.RedHat has been working hard at it for ten years and is only just a tiny niche still. RedHat ignored the desktop for many years because it was irrelevant in that market. Only this year did they get enough customers interested in FLOSS on the desktop to make it worthwhile to get back in. Dell, and HP, while having some investment in GNU/Linux actively promote that other OS. Against this, Ubuntu needs huge leverage to make a dent. Matt Asay may find his energy more useful in other markets which have much more flexibility and growth opportunities.

In his blog, Matt Asay has taken some strange views for an advocate of FLOSS:

• FLOSS is too cheap
• Record sales of “7″ is good for FLOSS
• Love/hate affair with GPL

Of course, I am paraphrasing his articles, but he looks at the same set of facts and comes to different conclusions. For example, the sales of “7″ show a record and I point out that much of that record came from deferred sales with upgrades from Vista in previous quarters while Matt Asay points out that M$ is taking the lion’s share of profit from the mouths of its partners and they may see GNU/Linux as a way to assert independence. His error is that M$ keeps their slaves hungry and liking it by pointing out the obvious fact that if they switched to GNU/Linux they wold lose huge market share and die. No OEM dares declare itself free of that other OS because the market for FLOSS is too small to sustain any one of them. They also cannot gradually shift to GNU/Linux for the same reason and more, for the $50 or so they give M$, they get $50 or more from the buyers of the PCs. If they sold the PCs with GNU/Linux or nothing they would make $50 less per PC. The OEMs are in a welfare trap. If all OEMs decided tomorrow to sell PCs with OS optional/separate things would change but no matter what Ubuntu does, that will not change until OEMs that do not promote that other OS step forward and embrace FLOSS. These are likely to be the smaller OEMs and startups, not the big guys who are stuck in M$’s pocket. It’s too bad we have to wait until GNU/Linux reaches a larger share of the market but that will happen even if the OEMs do not lead the way. When the market for GNU/Linux eats into their shares, they will promote GNU/Linux. Not before.

The GPL is what makes FLOSS work. Period. There is nothing wrong with the licence that folks trying to sidestep it would fix. Live with it Matt.

The world needs cheap IT and they can get it using mass production/Moore’s Law/FLOSS. A side effect of Free Software licences is that you cannot charge a high price for it because others can distribute the same stuff for less. This is a good thing, Matt. It means more affordable IT in emerging markets and greater innovation because the barriers to entry in any field of IT is less. Sell services, not FLOSS.

Change is good. We await signs that Matt Asay is good for Canonical/Ubuntu and whether Ubuntu is good for Matt Asay. New ways of doing business and new approaches to selling should be welcome. I only hope selling out is not in the cards. M$ has the cash and frequently buys out, embraces and extringuishes competition. It worries me that a new pragmatism may be creeping into Ubuntu. The FLOSS community accomplishes little by compromising on the principles of Free Software.

No, not Solitaire. Vista, the world’s most-hated OS. A man brought me a PC that could no longer dial out. It was a Conexant software modem and HP had a driver for it. All Vista would report was that the device “had a problem:”. Helpful, eh? The device manager did not even list a modem but called it a communications device. We downloaded the driver from HP but it failed to install with a cryptic message that HP states usually means nothing is wrong.

While we had the machine in the lab we did some updates to see whether that might be a problem. One service pack and dozens of updates later, there was no change. The thing now reported it had no driver and did not need one. We found fax modem software had been installed but was nowhere to be found. I asked the device manager to “scan for new devices” and suddenly the modem was there in the list all ready to go. I could disable and re-enable it, so I presume the driver does something. I wish on-line/HP chat help had suggested that, but they did not. They suggested uninstalling Norton, which was new and FireFox which was newer. It’s a good thing I wasn’t charging by the hour. The guy could have just handed over the machine as payment for the fiddling done.

This is an example of the high cost of using that other OS. Full of conveniences but costing much more than its price in downtime and labour. The world is not employed by M$ but does its bidding on hundreds of millions of machines. You should send M$ the bill.

HP has been a favourite company of mine for decades. In nuclear physics they made nearly indestructible power supplies, high-precision lab electronics and then computers. Now, I read on their site the usual stuff about recommending that other OS. Now, they trumpet “Do the Math”, recommending business adopt “7″ on new PCs to lower costs.

I can do the maths:

• it costs almost nothing to run GNU/Linux on the old machines - if they die, chuck them and use new thin clients at less than half the price of HP thin clients
• of course folks spending a ton of money keeping that other OS running may see “7″ on new as an option but they have far better options

HP knows about thin clients? They are the leading supplier. They know you can run them very nicely with GNU/Linux, yet they recommend that other OS. Does GM recommend Cadillacs? No. They make them for people who want an expensive car. It is silly to recommend the most expensive line for every customer. There may be some customers of HP for whom “7″ on new machines is the best choice but they must be in the minority or “7″ would be doing a lot more to pump up the PC industry. Instead “7″ is holding the PC industry back my putting a pricey roadblock on renewed IT.

This is the continuing soap-opera that is M$’s marketing schemes. They persuaded HP to put a damper on sales of GNU/Linux in 2002/2003 for a few shekels. When is HP going to learn that they should be working for themselves and their customers, not M$? The same could be asked of ASUS, MSI, Dell, and others who know there is a good market for GNU/Linux but they hold it back at M$’s behest. I think 2010 will end this nonsense. It’s about time.

More maths:

• $100 thin client
• $125 monitor
• $25 keyboard and optical mouse
• $25 worth of a good server

$275. That’s all the hardware will cost to put a good thin client seat on your LAN. The box could last you ten years, the monitor almost as long. Why are you spending $1000 per year per seat running that other OS on thick clients? You can add software for every client in seconds. You will rarely need to reboot. Students, secretaries, salespeople can move around and log into their sessions without losing their page. One person can manage thousands of seats. The secretary can plug in a new thin client/keyboard/mouse if one should fail. Why continue to bleed money for IT when there is much more that can be done for much less? Do the maths, please!

Stop recommending and accepting the recommendation of “7″ on new and shop around. You can find people who do not “recommend “7″”: IBM, RedHat, Novell, and hordes of others. Get off the wintel treadmill.

Not one of my favourite languages but very popular for rapid development of web applications, this interpreted language does a decent job on the front-end/ web facing/user-interface. However, it is interpreted which is extra work for the server. Caching can help but the answer is a compiler. Use PHP as it is to debug/develop rapidly your site but compile the code into machine readable stuff for faster execution. That is the obvious way to go and I did think to do such a project years ago, but it never amounted to anything. My concept was to rewrite important PHP scripts in Pascal so that they could compile properly.

The trolls here have pooh-poohed this idea, citing Facebook as an example of a high-powered/busy site running lots of PHP. Low, and behold, there is a story out that Facebook is working on a compiler for PHP.

“

Rumpus: So what will be the net effect of running the site on Hyper PHP?

Employee: We’re going to reduce our CPU usage on our servers by 80%, so practically, users will just see this as a faster site. Pages will load in one fifth of the time that they used to.

“

Facebook agrees. PHP is too slow as an interpreted language. QED.

Now, there are aspects of PHP that do not lend themselves to compiling like loose/flexible typing. The compiler has to make a choice at compile-time. This could be a fork of PHP or a new interface. We await the news.

Update: I tested the Roadsend PHP compiler against “Hello, World!”. My Pascal version executes in 0.001s according to time. The Roadsend version takes 0.1s. The difference? The statically linked version is 5.8MB! while my Pascal binary is 0.1MB. Maybe Hello, World! is not the best benchmark but it is an indication of problems with PHP compilers. Perhaps this is not an issue for FastCGI.

Update: Facebook has announced their project, HipHop,

“HipHop programmatically transforms your PHP source code into highly optimized C++ and then uses g++ to compile it. HipHop executes the source code in a semantically equivalent manner and sacrifices some rarely used features — such as eval() — in exchange for improved performance. HipHop includes a code transformer, a reimplementation of PHP’s runtime system, and a rewrite of many common PHP Extensions to take advantage of these performance optimizations.”

They have tweaked the language a bit and the runtime and compile it in C++ instead of interpreting the code. They get a 50% reduction in server load doing that so it pays in a big way. They will release the code as FLOSS later today. 50% may be no motivation at all for a light-weight site but for Facebook, it probably will save them millions annually in operating costs. For me, it could cut response time which is a great thing in a web app.