September 2007 Archives

2007-09-30 19:43:43

Bad luck with election advertisement

Some parties appear to have chosen their election advertisement rather badly. Those are namely the UDC, and the UDC.

The first bill they published is about some person they styled to look remotely like John Fitzgerald Kennedy. The bill has the title «The one with the Kennedy effect». People started questioning themselves whether this means that he attracts Marilyn Monroe or whether he has a bullet in the head.

Similarly well goes the incognito campaign «liste33.tv» of UDC. A black bill which only reads liste33.tv and «The average receipient of social welfare is in his early 30s, undereducated and a foreigner.» The bill then goes on to ask «Do you want this?»

Let's assume for a second that the above premise is true. In this case, the question whether or not one wants it is purely irrelevant. We are not talking about a vision in this case, but about a fact, so the question can be reduced to «Do you want to accept reality?»

However, there is also a problem with the validity of the premise. It can never be true. In Switzerland, foreigners are not eligible to receive social welfare. Thus, the conditions outlined can never be fulfilled.

As such it might seem pretty obvious why the UDC has launched this as an incognito campaign...


Posted by Tonnerre Lombard | Permanent link | File under: news, politics

2007-09-30 19:23:56

The search engines express their opinions on Censorship

Questioned about privacy laws and censorship, search engine providers can indeed give out weird messages. This became clear after an inquiry to Google, Microsoft and Yahoo about their search politics relating to China. All of these search engine providers have separate search engines for display and use in China which respect the local legal framework. However, this framework demands both censorship and reports on who searched what. People searching for keywords such as «democracy» are to be turned in by the search engine provider.

Google

Inquired about China, Google hands out a lenghty document typed up personally which explains that business pressure demands Google to operate in China. However, Google would not be allowed in China unless they implement the legal framework. The situation is called unsatisfactory, and a solution is said to be seeked but has to take place on an international, political level.

Also, Google says that the request to turn in people searching for certain keywords is not binding, and thus Google does not implement it.

Yahoo

Yahoo chose not to respond at all to inquiries related to China.

Microsoft

Questioned about China, Microsoft returned a prepared letter endorsing censorship as a perfect tool to keeping undesirable content away from the users. According to Microsoft, one should also look at the positive side of censorship. However, just like Google, Microsoft does not turn in users based on search requests.

Summary

It appears that Google is the only company which really has some kind of sense of corporate responsibility on the subject of censorship. It is however a fact that Google still plays with the dragons in this game, and hopefully Google will participate in any effort to clear up this issue in the future. The most unacceptable answer was probably that of Microsoft. Censorship is not acceptable under any circumstances, especially since it is not appropriate for any enterprise to decide on what a customer is willing to look at or not, for whatever reasons.


Posted by Tonnerre Lombard | Permanent link | File under: news, politics

2007-09-30 17:49:34

German "GI" endorses RFID and wants more

The german Gesellschaft für Informatik (Informatics Society) has held a meeting on transportation and logistics, promoting RFID due to its positive effects on that sector. Effects on the right to privacy remained unnaed.

According to the GI, up to 20 percent of the CO2 pollution caused by logistics could be eliminated using RFID tags in a distributed computing system. The GI imagines a routing protocol between the RFID chips to determine both message and physical paths from source to destination. This is what logistics companies already do using barcodes (packet tracking systems) to some extent.

According to the GI, both the tagged items and messages between the items shall be routed «like on the Internet». For physical routing of the tagged items, this shouldn't be a problem since it would be relatively easy to determine an «AS path» at the packing station, which would mean however that the routing information for the entire planet has to be kept at every packing station. However, aggregation is likely to happen in an adequately numbered system.

Inter-chip routing is however not feasible this way. There is no reason to assume that RFID chip numbers will be in any particular order at any possible point in time, so the average prefix size, assuming packages of 128 pieces are sent at average, would be 289, or 618'970'019'642'690'137'449'562'112, with 96-bit RFIDs. Aggregation would most likely be hard to impossible. But so is keeping a routing table of 619 septillion entries on every RFID chip in the world.

If every of these prefixes consisted of a 96-bit identifier and a 7-bit mask, the amount of memory to be kept on every RFID chip would be 492'581'209'243'648 Petabytes, or 481'036'337'152 Exabytes (481 billion exabytes). With Transflash technology, it is possible to store as much as 8GB on a square centimeter. Such a storage, which would have to save data without refresh cycles, i.e. battery power, for passive RFID chips, would thus have a size of 64'563'604'257'983'430'656 square centimeters, or 6'456'360'425 square kilometers (6 billion square kilometers). This is 673 times the area of China, or 12 times the area of the entire planet (including seas).

Also, privacy related aspects have been left out entirely. It would be desirable to also have a little focus on these issues which have been raised by the Stop RFID campaign. Otherwise, the science of informatics risks to fall to the Wernher von Braun excuse: «Once the rockets are up, who cares where they come down, that's not my department!» says Wernher von Braun.


Posted by Tonnerre Lombard | Permanent link | File under: general, news

2007-09-19 22:39:07

Making money with bank-to-bank transactions

Being annoyed again with the long delays it takes to transfer money between accounts at different banks. This becomes even more stupid with the new «transaction preview feature». If the transaction manager is able to determine when a transaction is going to arrive, and to mark this event in a timeline, why couldn't it just book it at that point in time?

The banks claim several things. Some claim that the delay, which is usually about 3 days, is caused by the fact that the requests are processed manually. However, they are usually processed on the first day, and usually this does not take place manually. Another popular excuse is that transactions are collected until some point in time and then synchronized. But that also proves itself bogus fairly simply. Try to create two transactions on two consecutive days: they will arrive on two consecutive days, rather than together.

The real reason is much more simple and not technical at all. For every day that the customer is in possession of the money, the bank usually pays a certain amount of money. However, the possession of the money allows the bank to make profit from it. Thus, the bank artificially introduces a delay before the money is transferred, but after it was removed from the account. Thus, the money is still disposable for profitable purposes, but does not cause any payments to the customer. This can still save a couple of millions of dollars.

Thus, both the incoming and the outgoing bank delay the transfer by one day each. More than that would cause riots along the customers, but one day is already a profitable margin for the additional ressource use. From a purely technical point of view, including accumulation of transactions, the maximum transfer time would be one day at maximum.


Posted by Tonnerre Lombard | Permanent link | File under: general, network

2007-09-10 20:49:36

UDC expels the expellation petition

The UDC campaign for a safer switzerland, demanding expellation of foreigners to their countries of origin, is actually hosted on a 30-day trial Typo3 hosting site.

And as if this wasn't bad enough yet, the trial typo3 hosting provider does not locate its servers in Switzerland – no, it is hosted on servers of a hosting provider in Germany.

What does this say about the credibility of the campaign?

Note: Yes, this is old news, but I didn't take the time to write it down yet.


Posted by Tonnerre Lombard | Permanent link | File under: news, network

2007-09-10 20:00:11

DE TLD moving to a new data center

DeNIC, the registry of the german top level domain DE, has announced to move the servers that keep the database for the TLD to a new datacenter. According to DeNIC, the old datacenter only offered limited potential for an additional growth of the namespace, which is expected by experts.

According to an article on Heise News, around 11 million customers will be affected tomorrow. Outages are not to be expected, DeNIC claims; however, as an expect, I would suggest that it's always better to consider them as potential.


Posted by Tonnerre Lombard | Permanent link | File under: news, network

2007-09-10 18:58:51

IBM joins the OpenOffice.org community

IBM has announced today to join the OpenOffice.org community. As the first contribution into the OpenOffice.org codebase, parts of Lotus Notes are expected to be merged. Also, IBM is planning measures to improve the quality of the OpenOffice.org code base. Also, IBM expects to improve the integration of OpenOffice.org with its own products, which would lead to improved usability.

On the OpenOffice.org website, there is a full press release on the new alliance as well as a collection of Frequently Asked Questions.


Posted by Tonnerre Lombard | Permanent link | File under: news, standards

2007-09-10 00:55:15

When portability is not an issue

Onitake, a MacOS X user from Chaostreff Basel, has recently sent in a bug report to the authors of curl-loader, a HTTP load generator.

In this bug report, Onitake complains about the lack of portability in curl, which uses the glibc<2.4-only feature of a definition of PAGE_SIZE in a system header named <asm/page.h>. The POSIX base specification issue 6 (xpg6) defines a portable interface named sysconf to access system information such as the page size. Additionally, the definition was removed from <asm/page.h> in recent versions of the GNU libc, making this definition a no-go for the future.

In this light, the response of the developers seems even more blatant:

 1. When asm/page.g is not defined there is a custom definition of
 PAGE_SIZE.

 2. We do not care about neither non-linux platforms, nor about POSIX.

 Thanks, however, for your caring and attention.
 Closed,
 Robert Iakobashvili

A candidate for is-broken.ch?


Posted by Tonnerre Lombard | Permanent link | File under: programming, standards

2007-09-01 18:16:21

Media coverage on OOXML proceedings (2)

More details about tricks in processing the standard proposal of OOXML (ISO/IEC DIS 29500) have come to attention of several news agencies. The swedish technical newspaper «Ny Teknik» reports that Microsoft admits to have subverted the vote in Sweden. Netzpolitik.org even writes about selling out votes and done deals in the swedish standardization process.

Groklaw has a whole collection about articles: Norway and OOXML, Sweden's new abstention and France still looking for consensus in the actual vote.

Another source of lots of news is ZDnet. They interviewed Georg Greve of the Free Software Foundation, and have a summary of all technical problems with the standard. In their conclusion they give the standard bad marks.

The progressive swiss german news paper WOZ Die Wochzeitung has an article summarizing the problems with OOXML and the procedural issues that have been raised by independent experts. Similarly, the german Computerwoche reports about the manipulations.

Even the Wall Street Journal has discovered the issue and is reporting on the committee stuffing and about Microsoft accusations against IBM to be the driving force behind ODF.

With all these interesting news, we all are looking forward to the ISO decision.


Posted by Tonnerre Lombard | Permanent link | File under: news, standards