Thursday, December 31, 2009

Romania's unsafe roads

I have to write again about the roads here.
Apart from being incredibly low speed designed - as they are going in a proportion of 40% through the villages (I am talking here about major European roads as E69, E70, E671, ...) they are in an incredible bad state. They are so full of holes that they look as a they have been made of Schweitzer cheese loafs. 400 km in Romania means about 8 hours so an average speed of about 50 km/h. In the cities the lines drawn on the roads are almost in any cases invisible so when it is to head for a certain direction you get a little confused. The drivers are carelessly stopping on the rightmost (first) lane so it is impossible to correctly drive on the tow lane roads as yoou are constantly obliged to change the lane and then to go back because on the second one the urban speeders are constantly flashing and horning.
It is exhausting to drive here....

Tuesday, December 29, 2009

Avatar dissappointment

Everybody praised the Avatar movie.
From my point of view the plot is quite thin while the CGI isnot as good as all the reviews said.
They said that the movement is very credible and the characters are realistic. Oh, no, the faces are absolutely horrible and the physics of the flying creatures is quite idiotic.
The landscapes are quite okay.
I think that LOTR had better CGI than this.

Saturday, December 12, 2009

Google everywhere

Google search, Google mail, Google wave, Google voice, Blogger, Google maps, Picasa.
I like those applications and I often use them. They are nice and clean and especially easy to use.
Then came the desktop applications Google Talk, Google Earth, Picasa, Widgets and the mighty Chrome browser. Again they are quite good desktop applications. But what I love most at Google is their development frameworks. GWT, Protocol buffers, Guice, etc. I am doing part of my job using them and they are awesome.
However there are two things I do not like at all about Google:

1. Go! language. It doesn't feel like a system programming language as it has quite an awkward syntax and misses pointer arithmetics.
And... What code does the code generator of this language creates? I remember that one of the first backdoors on Unix was to hack the C compiler to insert backdoor code for the login program whenever it compiles another C compiler... What if G does the same? I really do not trust this.

2. ChromeOS. I do not want google to control the internals of my OS and know everything about what I have on my hard drive. It is like having a Google Desktop search bolt into the OS itself. Ant it cannot be uninstalled or even disabled. Also the Google sign-on is strange. It means that I have to be online to use my device?

Tuesday, December 8, 2009

Perfect for today in Romania

Sonnet LXVI

Tired with all these, for restful death I cry,
As, to behold desert a beggar born,
And needy nothing trimm'd in jollity,
And purest faith unhappily forsworn,
And guilded honour shamefully misplaced,
And maiden virtue rudely strumpeted,
And right perfection wrongfully disgraced,
And strength by limping sway disabled,
And art made tongue-tied by authority,
And folly doctor-like controlling skill,
And simple truth miscall'd simplicity,
And captive good attending captain ill:
Tired with all these, from these would I be gone,
Save that, to die, I leave my love alone.

Monday, December 7, 2009

Projects wishlist

1. Python scripted SS7 simulator
A small simulator written in Python that could simulate ISUP/TCAP/INAP. Every message could be a class. Simple integration into automated tests via Python's built in 'unittest' module. Easy to learn for telecom people as Python syntax is very simple.
Something SWIG based for message wrapping?

2. Java ISUP interfaces
For easier integration into JSLEE environments.
Javolution based?

3. IMS virtualization
Quite simple to do - only hardware is the problem. I'd do an installation of OpenIMS Core on several virtual machines and then play a little bit with them. I'd check if Ericsson SDS can be used with this.

4. Statistics processor
Java application that should read statistics tickets written in a LISP like syntax and then store them in a database and display them out in different graphical/textual formats.

Sunday, December 6, 2009

Exit poll (part II)

The results are now reversed.
This is a confirmation that Circus will reign for the next five years here.
I mind asking what was in Geoana's head to proclaim its victory before having the official results.
And I am very amused that all the comedy shows presented sketches with Geoana as a winner (maybe they had alternative versions and presented the wrong one).
There are only two possible explanations for this:
1. Basescu and his party forged the elections.
2. We were victims of a massive media intoxication last night.

Exit poll!

The political life in Romania went from bad to worse.
Mircea Geoana, in my opinion a stupid and non charismatic figure, won the elections with the help of Crin Antonescu. Not longer that two weeks ago Crin Antonescu hit hardly on Geoana but now they are friends. In fact Geoana did not win elections. Media won the elections. Geoana was a pussy that wah unable to sustain its ideas and even in the last confrontation he came with his homeworks not done.
On the other hand, Basescu lost on his own hand. Although he had powerful PR (especially online) he made lots of image mistakes during his mandate. The fatal one was that he fought against the most important media groups. Associations with Udrea and Ridzi, scandals with his daughters destroyed his image, an image that started with a 'wind of change'. Unfortunately he turned into an autocratic figure and he tried to manipulate (and in several occasions he really succeeded) the electors he started a war against the parliament and constitution. So having so many enemies there's no wonder that he lost.
But still, the results are not final...
I'll keep looking for opportunities elsewhere...

Saturday, December 5, 2009


One of the things it bothers me most is that most of the bugs I am working on are not functional ones. They are rather due to the fact that the integration teams that are using our software and are creating the products from them have a poor understanding of how packaging works, what are the protocols used between internal and external interfaces and generally refuse to think outside the bos as they stick in having a set of procedures to that they adhere strictly. They seem to like the army-like environment where the creativity and research are not welcomed.
As for ops teams - it is even worse.

Virtualization and stuff

This week was quite crappy from a technical point of view.
The VXML Interpreter we have been developing for some time now has to discuss with Nuance software for TTS/ASR.
The problem with Nuance is that it runs on x86 hardware. Our baseline is HP DL380 G6 running RHEL 5 which is purely x64. I had to make lots of tricks and combinations between Realspeak, Recognizer, SpeechServer and some compatibility libraries in order to have a working setup.
The backup solution was to create a RHEL4 + Nuance virtual machine and force the RHEL5 host to boot it and present it to the TTS+ASR client. Management was very reluctant to it. However this solution would have bring other advantages - the developers could have used the virtual machines for unitary testing.
In the meanwhile my SS7 simulator ceased to work. It was not a big issue - a DHCP conflict in the network - bu this determined me to have a look on the system. I have discovered that the whole simulator is in fact a system having two ATCA blades. One is a DHCP/BOOTP/TFTP server (hence the conflict) ant the other is a diskless blade with SS7 and SIGTRAN interfaces. Again both machines could be virtualized - so I could replicate the expensive simulation setup on every PC on my team. And all this at 0 costs.