Alas, A Blog!

Moving to Live Journal

I have decided to move to livejournal for my blog. I have migrated all of the contents from here to that site. Please update your bookmarks.
http://www.livejournal.com/users/harishpillay/

Wow, my name is in the New York Times' Firefox ad!

Kinda feels strange that I managed to get my name (and some of my colleagues from Red Hat) in the Firefox ad on December 15th.

How I wish IPOS will take the high road of being fair and state all the options available to people with the changes in the law. Yes, again, in today's Straits Times, there is an ad by IPOS about the changes in the law and not a single mention about the freely licensed open source solutions!

I still think that the IPOS is a tax-payer funded organization but they way they seem to be disinterested in putting out the messages about FOSS makes me wonder if they have special FUD^H^H^Hmarketing funds from you-know-who :-).

The Problem with "Research Companies"

I attended an event organized by a well known tech news group and sponsored by the company I am currently employed with on Tuesday, Dec 8. The event was kicked off by a speaker from a "research company" - you know, one of those that everyone wants to quote from as if they are authorative.

Anyway, the event was focussed on what the next big thing is and had Open Source as the theme. One would have thought that the research company would have something substantial to say. But, what the speaker had to say was just regurgitating stuff from some "research assumptions" about how people would use open source stuff.

A clear sign that it was going nowhere was when the speaker put up a slide that talked about how Microsoft is responding to open source. In that slide, there was a bunch of checked-off boxes saying that MS has responded and the 1st one said "Shared Source". Shared Source? As an equivalent response to Open Source? I had to ask the question at the question time and I managed to get the speaker to acknowledge that shared source is not at all like open source and that it was not to be concluded from the slide that they were the same. Imagine if I did not ask the question - there would have been people in the audience thinking that since this "research organization" person said it, it must be true.

The speaker also made a lot of statements that reflected the speaker's bias. It was said that Microsoft sees a lot of attacks and security vulnerabilities because if has the largest share of the desktop OS market (and the speaker conveniently left out mentioning that the open source webserver, Apache with almost 70% of the market is no where as attacked and vulnerable as MS' IIS). And, get this, the speaker went on to say that the people doing the attacks are probably from the Linux and open source worlds! When this remark was made, I really wanted to standup and heckle the speaker (but I had to restrain myself). The speaker also said that people deploying open source should be wary of copyright/patent suits from Microsoft!

What a moron! At the tea break, I polled the attendees and they were saying that the "research company" speaker was totally off the track and was a waste of time to sit and listen to.

These "research companies" must be held to a higher standard of honesty and accountability. I have never come across a "research company" that had anything credible or worthwhile. All of them claim that their research techniques are credible and "industry standard". If "industry standard" is humbug (which it is), then all of these chaps are just milking money from a whole lot of BS.

It would be good exercise to see how these "research companies" have fared over the years. If there is someone who has some report from say 1997 about how the IT industry will be 5 years out, it will be good to see how accurately (NOT!) the report predicted things. Why is it that some companies balk at seeing a fortune teller but are ok to pay for a $5000 report from one of these "research companies"?

May be I have to set up a Pillay Research company and do the same!

Oh, it was interesting to see a person from the BSA in the audience. I am sure the BSA rep was there to get some bullet points from the "research company" for more FUD from them.

What's wrong with Microsoft?

People have asked me why is it that I have a problem with Microsoft, for afterall, I did work in MS in the late 1980s.

It is really quite simple. Time for a small story. I first started using PCs when I was working in Automated Systems PL (now known as CSA Automated PL). That was in late 83/early 84. We had taken on the distributorship of Corona PCs. As any good lore needs, the machines were really state of the art of that time - 8088-based, 4.77MHz, 64K or 128K bytes of RAM, 5.25 inch floppy drives (2) and for their higher end machines, a 10MB full height harddisk, green or b/w screen.

And life was good.

I quickly became the inhouse guru of these PCs. It was realy fun - having to haul a Corona luggable machine (boy those were damn heavy) home was a privilege.

It was running MS DOS 1.1 and then became MS DOS 2.0 which suppoted directories. Wow, stuff from the Unix world which I was dabbling with as well running Venix for a 8088 machine as well. Actually, I am not sure if it was a 8088 machine or an early 80286. What impressed me about Venix was that it was multi-user. That made sense to me because we did not have machines on our desks (individual use machines were not there anyway).

Let me fast forward to 1988 when I was offered an position in MS under their co-op student program (I took a break from graduate school at Oregon State University). I did not plan on working in MS, but when it happened, it was a thrilling feeling - here I am being able to work for a company that was defining the use of technology.

I was part of the Operating Environment (openv) which was providing technical support for MS and PC DOS, Windows 286/286, MS OS/2 and the Windows 286/386 SDK (and indirectly the MS C compiler needed for the SDK).

My first office was at MS in Redmond (building 12 if I recall correctly), but we got relocated to Bellevue (Bellevue Plaza I think) becase of space constraints.

As part of the package, I was provided with a PC, a 80286 machine with 512K RAM, a 40M harddisk and a floppy drive. We had a monochrome monitor with it (connected via a Hercules hard). Some of my colleagues decided to upgrade the machine (we had to pay for it ourselves), and we bought a VEGA VGA card and a NEC MultiSync colour monitor. Running stuff in colour made more sense as it made tech support much better. Thinking back, I am wondering why MS did not pay for that upgrade - I think we had access to a colour workstation but that was not to be taken home.

Anyway, providing tech support to the people calling in about the Windows 286/386 SDK and the OS/2 Presentation Manager was very challenging to say the least.

Learning event-based coding was the first major change for me - considering that all my previous (upto 1988) coding was either in assembly, C or Pascal on a DOS based platform. The closest to event-based programming was in the device drivers and ISR (install but stay resident) types of coding I did. The fact that there was an event queue being managed by the "OS" was kinda new to me. Even the coding I did on some AT&T 3B2 UNIX machines were standard stuff.

The whole point of this history is to lead to the way we had to provide support to ISVs. When PSS (Product Support Services) gets a call - and it was not a toll free number - we are able to look up answers in the Knowledge Base, via a text-based vt100 terminal emulation - yes, we had to telnet into a VMS (I think). [Boy all of this writing up is bringing back some fond memories!]

When kb returns with answers to a question, you then are able to read off from the screen and determine if that helps the caller. While it was good that the kb was there to help, there were - perhaps 3 or 4 out of 10 times, a section on the kb article that was entitled "Microsoft Confidential". What that meant was that anything following that heading was not to be shared/revealed to the caller. Inevitably, it would contain something like the information provided in the kb might be invalid in a future release of the product or that it is an undocumented aspect or something like that. But what was bothersome was that I could not tell the caller of this caveat as MS wanted to keep it secret. Now you know why you will find books entitled like "Undocumented Windows Function Calls" etc. MS wanted to keep some of the APIs, function calls for itself and not share with people whose applications were helping sell windows or os/2!

This level of dishonesty was bothering me and this was being done when MS was still an underdog with the big bad guy being IBM.

For what it's worth, I decided to leave MS even though I was contemplating working in their LAN Development group. That group thought that TCP/IP is a university research project and the way you should do networking is using NetBIOS/NetBEUI!!

So, there in lies the seeds of how I got to a point of not trusting MS in what it does. The sad thing is that there are great people in the company - coders, testers, etc etc etc. And, yes, the money could have been good, but I think my conscience had the upper hand!

Over the next 10-15 years (1990 onwards), MS continued with it's predatory practises in technology by acquiring companies perceived as a threat and then either killing off the technology or repackaging it. An example of the latter is Visual Basic from Cooper Software. I thought VB was cool - it helped people develop apps quite easily for an event-driven enviroment, but the way MS has "extended" it without thought to security is where MS has failed.

I hope that I can begin to like MS. Heck, some of my good friends work there. I hope that MS will do the following:

Document and publish all the document and network standards of their products - word, ppt, excel, smb to name just 4.
Stop charging the Microsoft Tax on the OEMs when they ship PCs from the factories. Let the OEMs ship what the customers want in the computers and pay the MS tax only for those that have MS products installed.
You may keep your application code closed and proprietary, but let's compete on implementation. If you want to open source the code, be honest about it and make it fully OSI compliant - not the sham that is "shared source". Shared Source is a lie - see no touch. It seems to fool some countries and governments, but not for long.

So, I do not really hate Microsoft - I just do not agree with their business practises and principles. They can become the good guy quickly too. They just have to try to be honest.

In the meantime, I will do all that it takes to make sure that everyone I can cajole and influence understands the issues at stake and consider running Linux and open sourced products.

KA9Q - what a awesome piece of engineering!

Back when I was in MS, I was trying to find out how best I could add TCP/IP to the computer I had - we had LANMAN but that was it. No TCP/IP to the outside world. My "internet" email was t-harish@microsoft.uucp. Yes, it was a UUCP link to UUNET.

Sometime in 1987/88, I came across the KA9Q TCP/IP code that was used to allow ham radio operators to use the AX.25 network to do TCP/IP. What an awesome piece of engineering - the code was fully open sourced but copyrighted by Phil Karn KA9Q - but gave a very good insight into what was possible even on lowly 8088/80286 machines.

I was just looking to see if KA9Q is still active on the net, and came across this.

Microsoft screws up again!

Ballmer said last week in Singapore that Linux users will face suites for patent infringements in the kernel etc. As predicted, his own PR had to issue a correction (although he did actually mouth those words - CNBC Singapore carried the clip), saying that he did not actually say it.

Well, the interesting thing is that Ballmer is using the number - to begin with it was wrong - that was mentioned by Ravicher which was part of a study Ravicher did. If you read that link, you will see that Ravicher is saying that there could be potential violation and that is about it. Ballmer, for purposes of keeping up the FUD, has chosen to regurgitate the number without engaging any thought.

Indemification is a red herring!

I think this is perhaps the clearest article that debunks the need to indemnify the user against suits.

Why does he still put his foot in his mouth?

Steve Baller said yesterday that Linux users can be sued for patent infrigement etc etc. And that Asian governments could be the target. Why does he say things like that, and yet be able to sleep at night? Does he not hear himself and does he not realize how really stupid he is sounding?

But, his PR folks think he is a dork (looks like it is offline now).

Thanks to Microsoft, Linux Flies!

Talks about how Microsoft over the last few months has become the biggest champion os Linux. They have paid out US$2 billion to their rivals, Novell and Sun, only for them to in turn use the windfall to go after MS.

Some of these twists and turns are so mind-boggling.

Solaris 10 - Open Sourced?

These are very, very exciting times we live in. A formerly closed and proprietary operating system, Solaris, has been open sourced under what I think will be a GPL license.

Wow!

Now we should look at the great engineering in that code base and merge it into Linux (perferrably Fedora and then into Red Hat Enterprise Linux).

Why?

We do not need to yet another variant of the Unix-like operating system around. We have the *BSD stuff already and most, if not all, of the good ideas and engineering has been adopted from *BSD into Linux. Let's just continue with that tradition and do the same with Solaris.

It is mind-boggling to think how the Operating Systems world has changed and all of it being driven by the availability of a world-class open source operating system, Linux.

We are already seeing the bridging of the information gap between the haves and have-nots in countries like India, China, Vietnam, Brazil, Peru, nations in Africa etc. This *will not* have happened if the closed, proprietary software model that is being bandied around by the likes of Microsoft.

RMS in Singapore

I was invited by the Singapore Management University to be a participant at a talk by Richard M Stallman of the Free Software Foundation. It was held on November 1st at the SMU Auditorium.

RMS spoke about Free Software and Patents though he did say that it is a very wide set topic that will take a good two hours to cover.

I have to agree that software patents are inherently and that we must guard against this being instituted in Singapore. Alas, the Intellectual Property Office of Singapore does not have anything specific to this and I guess it is just as well.

Something to note about this talk - the session following RMS was to be a response to what he had to say and there were to be two persons - one myself and the other someone from the IDA.

What is very telling about the ambivalence of the Singapore government to Free Software (and by extension Open Source) is that the IDA representative "choose not to participate" in the discussion even though his name was advertised everywhere.

Ah, the beauty of a closed mind.

No redundancy in the gas supply to the power supply?

Why is it that it is only *after* a power failure that the powers-that-be realized that the way the natural gas was being supplied to the power stations here did not have buffers and redundancies? We are dealing with gas here, not diesel.

The tech industry in 2010

Lately, I have been wondering how the tech industry will look in just 6 years from now. My sense is that we will see a disaggregation of the way software is developed with a very large chunk of software and services commoditized and fully open sourced. Why? Because it makes economic sense. As the silicon eco-system gets to be more and more powerful, the ability of these systems to be useful for far more things will be a given.

Therefore, what is the value of all of these "n-tier" software architectures etc when a lot of things can be done with faster and better hardware?