The phantasm of security

After every terror incident out sand carving facility goes in an overdrive of security which lasts for a few weeks at the most followed by business as usual. The security measures implemented as a result of paranoia at best serve to create a false sense of security. At worst they assist the terrorist in achieving their goal.

Checking of identity cards & Decal's on the vehicle

While on the surface this is a very good move as it only allows authorized personel to enter the premises but, This measure has reduced the process of authorization to possession of a peice of plastic. Unless the identification card is checked against a pre-validated database this measure is useless. So the guard at the gate should have a PDA with a barcode/rfid reader to validate the badge and verify wether the person presenting the badge is the same as the person to whom the badge was issues.
Failing this simple step of security exposes us to the following risks

1. An unauthorized person gaining enterence to your premises by
1. Taking a snapshot of a valid badge from a distance and replicating the design and colors offline
2. Arranging for a small accident for a regular employee snatching his badge replacing teh photo and voila. full access to your premises.
3. 
   


2. Similarly on the decal side unless we validate the vehicle being teh same for which the decal was issues and the driver is an pre-authorized person. we just create another hole in our security fabric.
   

If we go ahead and ensure that our security check at the gate is proper and until and unless a person is properly validated he is not allowed to enter the premises, We have just managed to create another risk as now a gunman can just arrive at the company gates at 9 am when the entire staff is waiting at the gates to be frisked and take his pick of whom he wants to dispose.

The problem with MMORPG

Most role playing games (RPG) be they online or offline involve a lot of grinding. A classic example is runescape where you spend hours killing the same set of cows/chickens/imps/goblins for different 'jobs'/'quests'

The strategy for a player soon reduces to
do `X` until Points='Y'

In such a scenario the wise player usually resorts to an bot to perform the mundane tasks in the game while spending his time on the `more exciting` quests.

And it is not unusual to find a different game erupting between the game designers and the bot designers/users where the first party tries to prevent the bots from playing their games and the second party tries to remove the artificial roadblocks.

This resorting to bots represents a flaw in the game design. What it indicates is that the player who has bought your game finds its so boring that he has delegated the task of playing your game to another program......

The response of the game developer to the bots is to ban the player and/or change their code to detect/disable the bot.

Why not allow and encourage bots? Encourage your players to disassemble and study each others bots and find out a way to beat them?
If I am anyway going to let my bot run for 24 hrs and spend 30 minutes per day playing the game, Why not allow my bot to run on your servers send me a report of the result once per day And I will tweak my bot to perform better the next day. Saving both your bandwidth costs and developer cost.

The cost of designing IP

1. Average Salary of an engineer $70K - $120K Each for a minimum of 10 Engineers
   
2. Tool Licenses- free to a few million $
   
1. Simulator
2. Verification
3. Validation
4. Synthesis
5. DFT
6. Layout


3. Design Period 1 Year+
   

C and ESL

Basically A method of synthesizing C code..... On the surface it looks good as "any c programmer" can now "do Hardware" but are we at a stage where we can tolerate silicon bloatware? Todays tools are capable of experimenting with various implementation for a given C code..... But do they settle for a Local or A global minima?

My foray in the ESL domain.

My foray in the ESL domain started way back in 2002 when I spent some time dabbling with perl based code generators. For my effort I got a couple of scripts which took the design description in various tabular formats and generated the required RTL be it for Register file generation, IO/ Pin muxing or Top level integration . A Year or two LAter I started seeing startup's offering the same scripts in a better/worse packaging.

A few years later I found myselves designing generic modeling frameworks in systemC and exploring other alternatives like opnet/omnet/ns2 etc.
Later on I spent a good 3 years working on a complex IP design using Bluespec System Verilog. Overall It was a good learning experience and I learned a lot about design from bluespec.

Recently I spent a year working on template based design of a multi- million gate SoC. Again a very good learning experience... With Great potential to improve the design process.