Why?

As usual, J.D. "Illiad" Frazer is right on top. But why ?

Sampada (ಸಂಪದ) Love Day!

"Giving love to projects" is a concept innovated by the GNOME Project. The idea is simple. You identify an open source project which you're likely to give more attention to and work on it. You love your work, the work automatically gets done.

Today, I'd a similar opportunity to give love to. The project called Sampada (ಸಂಪದ) is a community portal run by a very dynamic and brilliant friend of mine. The portal is hosted on a server at a US Data Center and used to run Debian GNU/Linux Sarge. It had a very good uptime and it was being managed well. Quite a few services are hosted on the server including the website - PHP, Apache, the database - MySQL, Email for the domain (Postfix/Courier/Amavis/Spamassassin/ClamAV) and a management interface - SysCP, Mailing lists using GNU/Mailman. All these were working steadily without any major hitches from some time, but we felt we needed to upgrade the whole system to the newly released Etch distribution.

Typically this kind of entire system upgrade entails a huge amount of planning, dry runs, down time loss estimations. In addition to lots of caffeine to release the tension caused by all the nail biting edges the experience puts one through.

In this case, we decided that we need to upgrade the server and we'd devote some time this Sunday and that's that. No more planning than saying "we'll get the thing done!". Cocky! IMO. But we were only dealing with the world's most stable and most advanced (In many ways) platform and we'd enough experience to be so lethargic about not making those painstaking plans.

We started a bit late, but we started (unlike other days).

# apt-get update && aptitude dist-upgrade -y

The first thing, we did was to ensure a backup of the most important parts of the disk was taken and stored offsite. Took us about 2 hours. Then the initial run itself went pretty smooth and we've had most of the system replaced by newer versions. But there were a couple of packages, which had some issues. Turned out that the customisation that were done on the configurations of those packages - Amavisd-New, Proftpd weren't compatible with the newer versions.

But that was soon solved by slight brute force. What we did was to specifically seek the version we wanted to install on the system. This happened in the case of PHP 5, Apache2. Soon, it was done. And when we tested, voila it worked straight.

1 hour straight. Some questions asked. All tougher questions parried, prayed to god and hoped for the best. But the whole upgrade process was so Debianish!! Always reliable! Always works!!

Overall, we now have a spiffy and snazzy new operating system and an ecosystem of programs serving out very interesting, intellectual and colloquial thoughts to anybody who simply seeks from any corner of the world. Don't miss out!!

ಹೊಸ ಚಿಗುರು, ಹಳೆ ಬೇರು
ಕೂಡಿರಲು ಮರ ಸೊಗಸು

ಮಧ್ಯೆ ಚಿಗುರು ಗಿಡವಾಗಿ
ಗಿಡವು ಮರವಾಗುತಿರಲು,
ಆ ಗಿಡ ಮರಗಳಿಗೆ
ನೀರುಣಿಸಿ ಬೆಳೆಯುವುದಾ
ನೋಡುವುದಿನ್ನೆಂಥಾ ಕನಸು - ಮಂಕುತಿಮ್ಮ

Truly Plural!

I was working on a small shell script to generate stubs for a project we're working on. After arriving at a rough design and deciding on the class names, method names to be generated, I started generating the stubs (scaffolding?). And came across an amusing non technical issue. getBench() was fine, but getBenchs() wasn't. It should've been getBenches().

Uh, oh! There might be a tool that can generate plurals given a word ?

Doing a quick search threw up a perl library to achieve the same. Three cheers for Perl !!! I'm always amazed at the huge number of modules Perl has. Anyway's here's a simple wrapper I wrote for that library which you can use to get plurals for most nouns.

shashi@anacoluthon:~$ pluralize.pl bench
benches

While at it, I happened to notice there's an implementation for Python and also a Grammar ruleset for implementing elsewhere.

PS: Title Inspiration: "A truly rural frugal ruler's mural" - tongue twister

Of starting up blues

Though technically not a startup, we've been in a startup mode from quite sometime - say seven years. Oops! In seven years, history gets changed by companies younger than that. So, what does it take to build a startup and be successful while at it ?

Marc Andreessen has addressed some points in his blog, in which he points out the reasons not to build startups!! He hits the nail right on the head with each and every letter in that post. Ouch! That hurts!! Let me try to correlate.

Why do we start up in the first place ?

• the opportunity to be in control of your own destiny
• The opportunity to create something new
• The opportunity to have an impact on the world
• create your ideal culture and work with a dream team of people
• money

The Opportunity to do. The Freedom to explore the Opportunities. The Opportunity to be Free. When compared to an engineering or an executive job, where you are one of the minions working to realise somebody else's ideas and opportunities, working for your own place under the sun is a very delicious idea. The freedom this portends, counters all the comforts that a cushy job can provide.

We've had numerous opportunities to explore over the years. Several ecsastic moments, several downfalls. The one big difference is that we didn't build to scale. We worked on the by now classical form of outsourced services. Having a replicable or assembly line model of products or services is going to make a big difference to the growth potential. We didn't do that!

Where can it go wrong ?

First, and most importantly, realize that a startup puts you on an emotional rollercoaster unlike anything you have ever experienced.

And what a rollercoaster!!! It ain't like nothing that can be experienced elsewhere. Further

You will flip rapidly from a day in which you are euphorically convinced you are going to own the world, to a day in which doom seems only weeks away and you feel completely ruined, and back again.

Know what??!! It need not be a day, even hours! minutes!! seconds!!! A positive side effect is going to be that you'll become a lot wiser, philosophical and equanimous. Of course, you need to have solid backing from your near and dear ones. And you need to have an outlet for your emotions without which you might end up in extremities. Like my friend, who killed himself a couple of weeks back :-(

In an established company -- no matter how poorly run or demoralized -- things happen. They just happen. People come in to work. Code gets written. User interfaces get designed. Servers get provisioned. Markets get analyzed. Pricing gets studied and determined. Sales calls get made. The wastebaskets get emptied. And so on.

Sigh! Those aren't something one person can do - day in and day out. Why didn't somebody tell me before ?

In a startup it is very easy for the code to not get written, for the user interfaces to not get designed... for people to not come into work... and for the wastebaskets to not get emptied.

You as the founder have to put all of these systems and routines and habits in place and get everyone actually rowing -- forget even about rowing in the right direction: just rowing at all is hard enough at the start.

And until you do, absolutely nothing happens.

Unless, of course, you do it yourself.

Have fun emptying those wastebaskets.

Thanks! Why am I pasting entire paragraphs here ? Because these are the exact thoughts I've been trying to put into words. Marc has articulated them so well. About the wastebaskets, "been there, done that!".

By that I mean that half or more of the people you hire aren't going to work out. They're going to be too lazy, too slow, easily rattled, political, bipolar, or psychotic.

Here, I've had quite a pleasant experience so far. Unless they are influenced by various factors, they do try hard, very hard to contribute. But, the issue remains that with all the paucity of resources, how is a startup going to manage it's people ? This according to me is the most complex challenge in the mix.

Fifth, God help you, at some point you're going to have to hire executives.

At what point do you hire executives ? And how do you compensate them ? Executives aren't cheap! They need to have a sense of belonging, if they're going to put in the effort. How do you address their hierarchy or position in the management ? What are the established roles in startups ? How do you shed your ego in order to work with someone with a more or less equal ego ?

Tough questions. If you've the answers, half the battle is won. The other half starts now.

Sixth, the hours
...
And even if you can help your employees have proper work/life balance, as a founder you certainly won't.

Stress. Strain. Pressure. And to think of it, I didn't think much of them even in my Engineering Mechanics classes. Maybe I'd flunked. Oh! wait :-p

It takes time for the culture of any company to become "set" -- for the team of people who have come together for the first time to decide collectively what they're all about, what they value -- and how they look at challenge and adversity.

In the best case, you get an amazing dynamic of people really pulling together, supporting one another, and working their collective tails off in pursuit of a dream.

In the worst case, you end up with widespread, self-reinforcing bitterness, disillusionment, cynicism, bad morale, contempt for management, and depression.

And you as the founder have much less influence over this than you'll think you do.

And by the time, a team settles. It's time for a new team to move in. Due to unlimited opportunities and these pressures, people move on. The camaraderie breaks. The work suffers. Customers get impatient :-)

Eighth, there are lots of X factors that can come along and whup you right upside the head, and there's absolutely nothing you can do about them.

Like our scheduled and unscheduled power cuts, lack of internet access, etc.

But, at the end of the day the thrill of coordinating all the activities, putting on various thinking hats, talking to people and getting things done is the bottom line for us - the founders/managers.

Parable of a MS Word document

Inspired by Mark Shuttleworth's latest post

A friend of mine is pretty tech savvy and also very organised. He backs up all his files once a month, copied it to floppies earlier, onto CDs later and now to DVDs. He carefully catalogs those disks and stores it away. Who knows, what might be required at a later day!!

He has in his collection all the documents, he's created over the years using a variety of tools - majorly Microsoft Office. He even has .doc documents that were created using Microsoft Word 3.0. And a lot of other spreadsheets (.xls) amongst others.

As any other tech savvy guy, he likes to keep his software always up-to-date. He has spent a fortune over the years on the Microsoft operating systems - DOS, Win 3.1, 95, 98, ME(!!), XP and now Vista !!! He has also bought all the Office versions right from the 3.0 days.

A great model customer for Microsoft. Nice! But so far, he hadn't had any use for his disks. He didn't want any of his documents from previous years. But, like any other intelligent investor, he still bought his insurance - made backups regularly.

Recently he'd upgraded his operating system from Microsoft Windows XP to Microsoft Windows Vista, and office from Microsoft Office XP to Microsoft Office 2007. As always, he muttered some profanities about absence of any compelling feature and that the program was stagnant from years now. Fine! He was ok with that.

Until last week! He was queried to provide a letter he'd sent to his customer some years back, in which he'd detailed the technical aspects of a work he'd done for them. They were happy with his work then and were looking for his services yet again. They wanted the letter to recommend him for the work, but had lost the letter and asked him for a copy of the same.

My much-organised-friend was more than happy, he looked up his catalogue - that he kept printed and filed away (:-P), checked out the disk which had the letter and popped it into his computer. The letter was typed using MS Word 95. But, it's all .doc right ? Yeah! Now the fun started. He opened the letter in his latest spiffy Office 2007.

It opened and what did he see ? Garbled text!!!!!!! He was shocked. He thought the disk was damaged. But it was a CDROM, well maintained and he could open the other files - .PDF, .TXT, .JPG, everything else without any problem. Hmmm! After some googling, he found out it was some unexplained problem that occurred with newer versions of MS Office. He also found out, this version of MS Office would not help him solve his problem.

He had a brainwave. He still had his copy of the Original MS Office 95 (TM) lying around. Why not install it and get it open it up. It'll take only a couple of minutes to do that right ? Wrong! He popped in the CD, clicked on Setup. It spewed out lots of complaints and aborted the installation. Huh!! He was stuck. I gave him a suggestion that he install his older copy of Windows 98 and try to install Office 95 on that :-P. He was horrified by the idea. How could he replace his latest Vista with some old buggy OS !!! And he would have to reinstall and redo all his customisations if he had to reinstall :-(

It was then, I asked him to install Open Office 2.0. He said, he didn't have the license. I said, I had the license and an unlimited license at that ;-) I downloaded the latest build for Windows, installed and there it was. His letter opened perfectly well, which he took a printout and sent it across.

All's well that ends well !! Not quite. This is not the end of the story. Proprietary formats like Microsoft's .doc, .xls lock up documents [ಕನ್ನಡ]. Our documents. Our pieces of work. Our data.

So, for this reason a bunch of guys(!!) got together and came up with a standard specification - Open Document Format (ODF). This specification has been implemented in tools like OpenOffice, StarOffice, Abiword and is being adopted internally by several corporates and by several Governments the world over. And as usual, Microsoft wouldn't agree. It came up with it's own "Open" format specification, which can at best be described as an "open" container for it's binary formulations.

Microsoft says, they have evolved the .doc format to such a level, now it can support all kind of media, supports versioning, is tied with several tools (Example), and hence the new .docx format should be the standard. Huh! People who are in the know have certain observations, which isn't necessarily nice for Microsoft. Microsoft says, they'll provide converters(!!) for various formats. I would rather write my own.

The advantages with Simple, Open and Featureful specifications such as ODF are that, it gives rise to several opportunities without having to depend upon it's creator. And the best part is that these formats are always open.

My data should remain mine alone, not some blood hound corporation's.

Fortune 100!!

In unix, fortune has been around from a long time and it's as popular and a favourite now, as it was way back in time.

Some nuggets:

Don't worry so loud, your roommate can't think.
----
You will wish you hadn't.
----
Don't tell any big lies today. Small ones can be just as effective.
----
You've been leading a dog's life. Stay off the furniture.
----
You will experience a strong urge to do good; but it will pass.
----
Be free and open and breezy! Enjoy! Things won't get any better so
get used to it.
----
You are a very redundant person, that's what kind of person you are.
----

Finally,

$ fortune
You have literary talent that you should take pains to develop.

How true :-)

Corporate Responsibility!?

Our honorable Prime Minister has called upon our corporate honchos to cut on their drawings and increase their commitment towards social responsibility. As several writers have already talked about it already, that's a very timely suggestion and it's about time too. Or is it ?

Indian corporates have always been humble over the years. They've also have had some ideas of social responsibility. But, nevertheless the government has been totally distrustful of the private sector and have taken it upon themselves to better the lot of the people.

Of course, there are several interpretations as how to improve the common man. But the government almost always has botched up every opportunity, except for only a few instances. Our Prime Minister realised this two decades back, and started involving the people (private) in the business of running the country. Unfortunately, some took it too seriously, some didn't and many others celebrated the return of cerebral politics.

This too was short lived. The government created partners in the form of middlemen, brokers, allowed every rule in the book become redundant and promoted the greatest aristocratic enterprises (read property dealers and cos). The government, in all it's wisdom procures land and allows the industry to do whatever they want, on the land. Even setup entertainment centers, recreation places (ok! it's needed. but on fertile farmland ?)

As the prime minister says "But unless the governance of the economy is improved, even further liberalization – which is sorely needed – will be insufficient to sustain growth". How is the governance going to improve ? Move it to the e-sphere ? eGovernance ? And how to adopt eGovernance ?
Just having a look at a concept papers and working papers from the government's standardization website - http://egovstandards.gov.in/ gives other ideas. Most of the papers speak at a very high level, just cite academic theories.

eGovernance in the Indian context is seemingly difficult. But if there's will there's a way to bull doze :-)
One of the finest implementations of eGovernance has been done in MCA - http://www.mca.gov.in/
It has improved the efficiency and has virtually zero scope for corruption in the processes.

"The world has enough to meet everyone's need, but not everyone's greed" - M.K.Gandhiji

Theory of Relative "Product"ivity!

A few weeks ago, Userfriendly ran this series:

So ? Google betas their products almost always, to enable people experience cutting edge technology without waiting endlessly for the product to pass all kinds of QA tests. In the process, the users are helping Google harness their resources much better and test the application while at it. Noble thoughts indeed!!

Aaargghh!! Here comes Marketing. But what if, it is a marketing ploy! There are hundreds of such alpha and beta products out there. What prevents one from launching a nice looking online dating site, not with much to show for, but some ideas and beautiful pictures on the screen or on the ad and brand it alpha/beta!!?


Companies do that to drum up interest in their site, get decent numbers on their websites, show the reports to interested VCs and Funds, raise capital, hire more people, repeat cycle. But the fun starts when GMail like fiascos happen.

Issues happen. Always!

But how ethical is it to sell to users to use a beta system, while there's every chance of the user's data being wiped out or the system not being available. Perhaps there should be some kind of browser plugins that keep warning the user this website is still in beta or even alpha.

Bottom Line:
Use sites that are in Beta/Alpha. Help Stef buy his Porsche.

Free Operating Systems. Next"enta" ?

enta (ಎಂತ) in Kannada slang translates to "whither" as in "whither are we wandering ?".

Over the past couple of months huge strides have been made in the development and deployments of several new free operating systems. While tracking their progress and goals, I'm writing down a couple of points for myself to revisit some time later and make an assessment.

Debian GNU/Linux: The universal operating system just shed a couple of architectures so that it is now that much less universal :-p. But nevertheless, with the "etch" release, they've managed to stick it to the Ubuntu folks and have regained the trust of thousands of Linux enthusiasts all over the globe. In addition to the main release with over 20,000 packages spanning at least 3 DVDs, Debian also has a thriving sub-projects culture which is progressing quite steadily. These sub-projects cover specialised applications including Embedded Debian, Debian Junior, Debian Edu, Debian Lex. All these positive vibes should take Debian beyond the hard core GNU users base it currently enjoys. But, the project has to overcome issues such as the Firefox/Iceweasel, GNU Free Documentation fiasco which can irritate neutral users no end.

Overall Debian is on it's way to nirvana and helping it's users attain nirvana.

Ubuntu Linux: The hot favourite amongst the middle class of operating system users, Ubuntu has also taken a step ahead to ensure it reaches the masses. Ubuntu is now available as an option with Dell. This is sure great news for customers, as hardware incompatibility and other issues might not occur. Ubuntu also did great with their 6.06 LTS release. We ourselves have deployed the same at many corporates and institutions. Ubuntu comes with semi-annual releases.

Red Hat Enterprise Editions and their forks: These are interesting distributions. Red Hat's Enterprise Linux is just that. It's enterpricey. It's performance, It's Ease of Use, It's cost. Everything is so enterpricey. Some good souls have come out to give us less privileged folks, a taste of the enterpricey experience at a very less or at no cost. These include CentOS, Whitebox Linux. But the irony of the whole linux in enteprice is that Oracle themselves do a repackaging and provide support for Red Hat EL binaries. Why do they do that ? Simple. Oracle needs a base. According to them, Oracle itself can tune the kernel, support programs they develop on oracle.

Fedora Linux: What started as Red Hat's staging area (??) is now a full blown enterprise capable distribution. It has a huge community built around who takes care of the packaging to the bug fixes, etc. Truly Red Hat has delivered it's word that Fedora won't be controlled by Red Hat.

Now the above Operating Systems were purely Linux based. What makes the scene much more interesting is that there are a couple of contenders for the growing (yet again) base of Unix users. These include Sun's OpenSolaris, the BSDs.

Sun has been doing some real marketing amongst developers out there to adopt OpenSolaris for their development purposes. They have made available cutting edge tools including DTrace, ZFS available under an open source license. There have been several instances where attempts to port these to Linux/BSD are going on furiously. But Sun isn't satisfied. The developers still do not trust Sun to play along nicely along open source rules. So, Sun has come up with something called Project Indiana. This has Ian Murdock (The Ian in Debian, Progeny founder) at it's helm. He looks like he has some tricks up his sleeve. Let's see if he can turn it around.

There's also this project called Nexenta, which is again a Debian based distribution but uses Solaris kernel at it's heart. And tries to recreate the Solaris environment.

Bottom line:
Projects, there are some
There are people using them
At office and home

Who does rule the roost
Will not necessarily be the best
But, try certainly will be doing the most

Hedging life !!??

Today, I'd been to a session where I was educated about the importance of hedging against the Dollar falling again against the Rupee. The gist was this, "you buy Futures the bank's gonna profit through margin money. You buy Options, the bank'll debit your fees". "You may lose some money but it's all about minimizing risk, honey!!" "Hedging is not about another way to increase your profit, but a way to minimize your losses".

Well! not exactly their words, but they meant it.

If it's so risky, WTF are we using Dollars to trade then ? "Then, Uncle (Sam) had heap many dollars. Now, Uncle borrows heap many dollars to pay us". If this news' anything to go by, it's a matter of time, countries switched to saner alternatives.

Now, what are the alternatives ?

Why not bill in Rupees ? Will the rupee hold ? Won't it fall if the boom is not sustenable ?
How 'bout the Euro ? Who says the Euro will not go the dollar way ?
Hmm. How 'bout gold ? Ha! Ha!! How are you going to bill ? One gold bar per hour :-)

So, the alternatives suck. And we need to hedge against the Dollar falling yet again. Or rising up!!
Whatever happens, the bank makes money!! Why not close shop and help others hedge as well ?

In Over the Hedge, RJ says:

RJ: Now, the traps are set here, here, here, here, here, here, here, here, here, here, and here. Here, here, here, here, big one here, here, and maybe a few over here.
Stella: Gee, it's that all?
RJ: No. There's bunch of red lights all over here. You OK, Verne? Look a little green.
Verne: I blacked out for a second there, but... I get the idea: there's lights, traps and I might need to change my shell.