March 11, 2011 at 12:48Bad dancer ...
- Transfer
Many of you have read about the latest Estonian elections. And here are some interesting facts.
The Estonian government has hired an IT vendor, Helmes, to create an operational data reporting system during the last Estonian parliamentary elections. They (Helmes) built a completely new system, and apparently did not test it in advance of the election. Unsurprisingly, the system crashed and statistics were submitted with an hour delay.
Why am I writing about this? Because Helmes Blames PostgreSQL Database Enginein delay. It’s the same, that the driver, after getting into a car accident, will blame the engine manufacturer, although he himself rushed to a red light. “If only the engine were a little more powerful,” Helmes complains, “we would have slipped through this damn intersection before the other cars start!”
Assuming Google Translate is adequate in its translation, Helmes provided a truly bizarre explanation for the lack of testing:
“The only way to prevent this situation would be to preload the data with the same amount of information as in the midst of the election. This is not normal, since the start of the system should not depend on any amount of pseudo-data. ”
In other words, Helmes never tested a system with a full database. The classic mistake of beginners.
PostgreSQL currently serves elections in Argentina, New South Wales (Australia), New Zealand, as well as several Brazilian states. The population of all these regions (well, maybe, except for New Zealand) is significantly larger than the population of Estonia, and no one in those places reported election failures due to performance problems. Hell, I saw SQLite-based selective systems, and they work great because they were designed correctly.
The same Estonian Skype handles more than 6% of long-distance calls in the world using PostgreSQL. This is about a billion transactions per day. Worldwide, PostgreSQL serves many systems that process volumes of several Estonian elections ... every hour, day after day.
My advice to the Estonian government: fire Helmes. You do not need a company that is not able to fulfill an order, and then blames its own tools for this.
If any of the readers know Estonian, please translate this message and share it with the people of Estonia.
(In addition, I hear MySQL apologists laughing at us now. Now we know how it is to be blamed because of the curvature of the hands of users of the system.)
The Estonian government has hired an IT vendor, Helmes, to create an operational data reporting system during the last Estonian parliamentary elections. They (Helmes) built a completely new system, and apparently did not test it in advance of the election. Unsurprisingly, the system crashed and statistics were submitted with an hour delay.
Why am I writing about this? Because Helmes Blames PostgreSQL Database Enginein delay. It’s the same, that the driver, after getting into a car accident, will blame the engine manufacturer, although he himself rushed to a red light. “If only the engine were a little more powerful,” Helmes complains, “we would have slipped through this damn intersection before the other cars start!”
Assuming Google Translate is adequate in its translation, Helmes provided a truly bizarre explanation for the lack of testing:
“The only way to prevent this situation would be to preload the data with the same amount of information as in the midst of the election. This is not normal, since the start of the system should not depend on any amount of pseudo-data. ”
In other words, Helmes never tested a system with a full database. The classic mistake of beginners.
PostgreSQL currently serves elections in Argentina, New South Wales (Australia), New Zealand, as well as several Brazilian states. The population of all these regions (well, maybe, except for New Zealand) is significantly larger than the population of Estonia, and no one in those places reported election failures due to performance problems. Hell, I saw SQLite-based selective systems, and they work great because they were designed correctly.
The same Estonian Skype handles more than 6% of long-distance calls in the world using PostgreSQL. This is about a billion transactions per day. Worldwide, PostgreSQL serves many systems that process volumes of several Estonian elections ... every hour, day after day.
My advice to the Estonian government: fire Helmes. You do not need a company that is not able to fulfill an order, and then blames its own tools for this.
If any of the readers know Estonian, please translate this message and share it with the people of Estonia.
(In addition, I hear MySQL apologists laughing at us now. Now we know how it is to be blamed because of the curvature of the hands of users of the system.)