Moving

Ok, so I've decided that there is really no point continuing this particular blog. For one thing, I dont really post much. Add to that the fact that the wordpress syndication with blogger (my blog atleast) doesnt really work by any standards, the fact that wordpress now has a much nicer interface, and the fact that I may as well post on my IITK blog instead for the next couple of years, and its pretty much the logical thing to do. I was holding on to this with the intention to keep this longer than the IITK one lasts (as long as I stay there, ie), but its now painfully clear to me that if by then I dont have the ability to host it on my own I should probably just stop blogging.

In case you are one of the few people who read this / are subscribed, I will henceforth be writing (or not writing, as the case may be) at http://blogs.iitk.ac.in/chintal/ . Its RSS feed is at http://blogs.iitk.ac.in/chintal/feed/ .

Note that this does not mean I will be posting more, just that I wont be posting here anymore - so dont hold your breath.

And they all look just the same

So I'm back in Hyderabad now. Its been a very hectic and interesting last two months, but now I get to relax for the next week or so before the next semester sets in. My life has changed a lot in the last three years, and it seems Hyderabad has too. It seems the generaly advertising scene in Hyderabad is predominantly real estate and housing related. By real estate I don't mean the typical buy-land-in-this-part-of-the-city kind. Instead, they all try to sell you 'luxury homes and bungalows' in so-and-so-part-of-the-city-that-you-never-heard-of-but-is-_the_-up-and-coming-neighbourhood-of-the-city. Even though I emphasize the part of the city they want you to pay truckloads of money to live in, that isnt the part that disturbs me the most. Its nice that the city is growing in many directions, and I sincerely hope the metro rail will come up in time to ease getting from A to B in the already sprawling city. The sad part, however, is that they all remind me of the title song to Weeds (a TV show that I very highly recommend). They all look just the same. Why would I want to pay in 8-digits (INR) and then live in a house which looks _exaclty_ the same as about 900 other houses in the same neighbourhood? One very important thing about luxury is uniqueness. One doesn't go to a 5- or 7-star hotel because it looks and feels exaclty at home, or (this is probably a better analogy) buy a ferrari to have a car that looks exactly like every other ferrari. They go to a posh hotel to feel _different_ from the usual, and buy a ferrari to look _different_ from every other car on the road. Its that shared uniqueness that would bring two ferrari owners together on the street. When was the last time you saw two guys who own a Maruti 800 or a Tata Indica pull over and exchange notes about how their cars behaved? To make things worse, its not just these 900 or so houses that look just the same. More or less _all_ the luxury bungalows on sale across the city look - and probably feel - pretty much the same. Sure, given the contraints of a typical house (so many rooms, the general configuration, then the superstitions dictating the placement of thest rooms, and so on) the number of possible permutations of these designs is very low to start with, but that doesnt mean that I'd be comfortable living in a house whose 'superior quality' requires me to bring out a tri-square and a spirit level to demonstrate. I can still remeber bit and pieces of the discussions that went in when my parents were building our house over 13 years ago... which way should so and so door open, how big should so and so thing be, so on. There are some really, really old houses I've been in where the architecture (and not just the age) tells a story. What we are doing now is very different from what was done years ago. Will these identical houses, over many decades, change and evolve with their owners over time? Will 50, 100 years down the line someone walk down that street where all the houses now are not just dimensionally identical but also the same color and texture and point to one house and say "I like that one. I don't know why, but it just seems like a better home than the one next to it"?

Plasma

I've been really busy with a whole bunch of stuff, and so dont have time to write about this, but I came across this image on PlanetKDE by Wade Jolson which I just wanted to post.



Just for the record, I haven't used KDE4 since I left campus (any my trusty desktop)... which was sometime in the end of May, but even at that point I was very happy with where Plasma was going. I can't wait to get back and get the latest svn version. I dont know why so many people have issues with the FolderView concept either ... that is exactly what I've really wanted for the last 4 or 5 years.

Help set a world record

EDIT :: The "official" Download Day (http://spreadfirefox.com/worldrecord)
will start at 1130 PM IST / 10am PDT / 1pm EDT / 6pm GMT / 7pm CET / 9pm MST

EDIT :: Firefox 3 Launch Date (and thus Download Day) is set at the 17th of June, 2008.

Most of the people who read this blog probably already know that the Firefox3 launch is coming up. You probably also know about the world record effort. In case you don't, then here's the deal - When Firefox2 came out, there were about 1.6 million downloads in the first 24 hours. The folks over Mozilla, SpreadFirefox, and a bunch of other groups in the community want to make the Firefox3 launch especially memorable by trying to push for a Guinness World Record. In order to participate, all you have to really do is download Firefox3, preferably from the Download Day Headquarters.
In the meanwhile :

• You can pledge to download Firefox3 on Download Day (the date for which is still in limbo, but it'll probably be sometime in the end of June), also at the headquarters.
  
• If you blog or have a website, you can put up download day buttons and banners and post about it.
• If you're one of the social networking types, there are groups/communities in orkut, facebook, myspace, et al. for you to join and use to send mass invitations for. (Just don't blame me if people such as myself but who don't care much for firefox in particular or freedom and good software in general bitch about being sent bulk invitations)
  
• You can follow firefox on twitter for updates.

There are some things to note, though, before you actually go and download :

• Please download only once per computer. Even if you do go through the trouble of downloading it a hundred times, not only is it not ethical, it will probably only count as one anyway.
  
• For a download to be counted it should be a complete, successful download. Updates don't count. When I installed the beta some weeks ago it did a pretty good job of bringing my settings and such over from the older install, so there probably won't be much complication in downloading the full installer and reinstalling instead of the update. The difference in bandwidth usage might not be very significant, either, if you are upgrading from Firefox 2.0.x .

As a side note, Firefox3 is a great improvement from Firefox2. I've been using it for quite a while now, and the RAM consumption issues have been reduced significantly. The interface is nice and clean (the defualt theme is, atleast). The settings panels have gone through some shuffling around again, but the one thats there now is noticably better than what I remember of Firefox2's settings, with more options and intuitive placement of stuff (although the proxy settings are still nicely hidden away, unfortunately). The improvements that take the cake for me are the new download manager, which _seems_ nice, although I can't put my finger on why, and the address/search bar. That one you have to use for a while to understand how much better it is than any other browser's address bar, even Firefox's old one.

As a second side note, here are the stats of the last 77 visitors to my blog on blogger. I admit the set of people who would visit my blog is rather biased to begin with, but its still nice to see these kind of numbers.

PIC, First Impressions.

Although I've had my share of fun with microcontrollers over the last few years, all of it was with AVR controllers, and most of that was with the ATMega16.

Almost 3 years ago, when I first joined the Robocon team, SPS had just started using the ATMega controllers, and we never looked back from then. I had heard of PICs, but never got around to using one, although its touted to be one of the easier microcontrollers to use. Some work I'm doing now had relatively strong reasons for using a PIC - a couple of PIC16F877As were already available, and more significantly, there was a programmer with it. The programmer is usually the bottleneck when it comes to first timers to the ATMega, and I figured since I had a commercial programmer at my disposal, it wouldnt be too much of a pain. I was wrong.

To start off with, searching on the internet for PIC related things seems to return results with an abnormally low signal to noise ratio. Even legitimate results are spread over a wide range of PICs, some now obsolete. Tons of pages are from many, many years ago and are no longer accurate. The pages that seem the most promising sometimes turn out to be AVR related pages.

PICs also seem to have massive dependance on various commercial toolchains. For an AVR, the avrgcc-avrdude toolchain is more or less enough to do a lot of things. There's really good documentation, and due to the fact that is open source, it has excellent community support channels.

SDCC is the only open source compiler that seems to support PICs, but its next to impossible to find sample code with explanations, or even halfway decent tutorials. To top it off, it seems SDCC is at best beta-quality when it comes to PICs at the moment. All the other compilers are commercial, and cost a pretty penny. Evaluation versions do exist, but I generally dont like using those unless there is not viable option, especially when they have restricted functionality. Eagle is one software for which I don't mind using the evaluation version (primarily because its way better than most of its competition), and I have on occasion used the CVAVR eval version when the rest of the team was not comfortable with avrgcc and we were on the clock. It would seem, however, that it is next to impossible to do stuff with the PIC without an endless stream of seemingly arbitrary proprietary software. Some of this software does look pretty nice, to be honest, but I'm still extremely pissed about how circuitous the path to get here has been. AVRs also have a lot of 3rd party compilers and IDEs and such available, but somehow avrgcc stands out in the crowd. Avrfreaks and the avr mailing lists seem to do a wonderful job of keeping things together, and there is always a somewhat authentic source to get information. SDCC for PICs still has a long, long way to get there. All of these issues came up on a windows machine. I have no idea how many hoops I'd have to jump through to get this done on a linux machine, considering the proprietary nature of the compiler, the linker, the programmer...

There are programmer circuits in the open domain, but theres a lot of noise there too. Try to search for in-circuit programming on PICs. It wasnt until I was able to track down the manual of my programmer that I was able to determine that it did, indeed, support ICSP. As for software that does the programming, there are tons of those too. If you dont know what programmer you have, you're going to spend a long time trying to figure out which software will do the job. For a person coming from the usual low-voltage ISP of the AVR family, where high voltage programming is something you do on the rare occasion when you need to fix some terrible mistake in fuse bits, the whole LVP disabling thing of the PIC is a little confusing, unless you're lucky enough to find some text describing the situation early on.

PIC programs also seem a little bit screwed up, with the configuration bits described in each program. Why those bits have to be reset on each program is something beyond me - on AVRs, fuse bits are something we change on rare occasion, and with much care. Although I must admit, the HVP on the PICs does make it easier to fix mistakes at that level.

I've also realized that the internal RC oscillator on the AVRs is also something that is often taken for granted. Thanks to the HVP and the way PICs get programmed, the programming works even when the chip isnt getting its clock properly. Again, thats nice to fix mistakes in setting the configuration bits, but somewhat a pain when you arent sure if your crystal is properly working with the capacitors and such.