Slashdot

I have been thinking a great deal about how Colombia can get out of the civil war which has torn the country apart for over four decades. The problems are deep but they may be soon solvable. In short, leadership of the government devoted to social justice, economic growth, and security for all Colombians is absolutely necessary. There is hope that this can happen but not under Alvarro Uribe.

Colombia's civil war began as a Communist insurrection in the 1960's. The goal of the Communist powers was probably to weaken US influence in the region especially in the wake of the decision by Colombia to send thousands of troops to fight in the Korean war. However, over time, FARC has cut all political ties to any political parties and has become primarily occupied with their own financial interests relating to drug trafficking. There are other leftist rebels in Colombia, but none of them match the force of the FARC.

In part to contain the FARC, the Colombian government has financed and sponsored a number of right-wing militias which are also into terrorism, narcotrafficking, etc. While not as big or as strong as the FARC, they are still a major force to be reckoned with. Unfortunately, the current president, Uribe, has continued policies of backing these militas. Consequently Colombians are left with no guarantees of security in a civil war where both sides readily resort to terrorism and where both sides finance their aims via the manufacture and sale of cocaine.

As hellish as the situation sounds (and Colombia is not likely to be somewhere I would go to visit at the moment), the beginnings of hope are starting to emerge. The peace process and FARC's handling of it have caused most of the left-wing of Colombian politics to cease supporting the organization, since their main goal has become that of narcotrafficking. While the right-wing and the government has not abandoned their terrorist organizations yet, political pressure is building to do so.

What Colombia needs is for a center-left candidate to emerge victorious in Presidential elections with a message that Colombia as a whole can unite behind. THe center-left part is important because this is necessary for being able to crebibly reject violence from the FARC. The message needs to be one of social justice, economic growth, and an attempt to provide security for all Colombians from the terrorist organizations which have dominated both sides of this conflict. Once Colombians turn away from violence, then the militias (including the FARC) can be taken down.

It will not be easy-- Uribe is seeking modifications to term limits to let him run again for the same office. In this regard, he joins the ranks of Hugo Chavez, Alberto Fujimori, and other Latin American authoritarian leaders who would rather rewrite the law than step down. While it is hoped that the measure doesn't pass, we will have to see. Secondly, any President able to marginalize the militias would almost certainly have a platform that the US would not like. There would be additional resistance to breaking the historic ties between the countries.

Nonetheless, I am hopefull it can be done. It seems possible that within another decade, this horrible civil war will be only a memory.

I have generally called Colombia's raid unbelievably stupid. However, I figured I would discuss who the real winners and losers are in the developing crisis:

Winner: The FARC. With this crisis, Colombia has recalled their troops away from border areas, giving FARC a large and safe corridor in which to operate. Venezuela's mobilization is also believed to give the leader of FARC military protection of the Venezuelan government. If FARC continues with negotiations and releases the hostages as previously expected, France has also expressed some willingness to cease regarding FARC as a terrorist organization. Legitimacy by even some EU members may be a strategic victory in the on-going conflict. (Personally, I think that FARC needs to be seen for what it is-- a large-scale mafia-like organized crime syndicate which does not recognize international borders. I am not sure I would call them "terrorists" so much as the "Colombian Mob." They have no real political platform other than their own economic interests in cocaine production.)

Winner: Ecuador. Ecuador has in the past taken a tough line against general FARC activity in Ecuador with the exception of offers to try to mediate a peace process and hostage release. The attack has been seen across Latin America as unacceptable, and Correa has gained much-needed support.

Winner: OAS. OAS has shown that they are capable of dealing with crises in the area and helping to get people to back away from the brink of war (Ecuador has threatened military retaliation against Colombia over the cross-border raid).

Loser: Hugo Chavez. Closing the border to Colombia will exacerbate inflation and food shortages in Venezuela and likely cause dissatisfaction in the longer run.

Loser: Averro Uribe. Uribe is standing alone in this crisis-- it is unclear how much support he has even in Colombia relating to his handling of the crisis. The economic toll to Colombia is likely to be a real problem in the ongoing civil war.

Winner: de Silva. Brazil has shown that they can be a real diplomatic powerhouse in the area. It is likely that Brazil's influence in South America will be strengthened by their role in this crisis.

Loser: The USA. Bush's handling of the crisis ensures without any doubt that the lease for the only USAF base in South America (in Manta, Ecuador) will not be renewed. The posturing of all presidential candidates on the issue have further weakened US credibility and diplomacy in the area. The leftist governments see the US as being imperial, while the right-wing governments see the US as undermining their sovereignty. This is not good for us.

To be fair, Colombia's actions would be like having the US, without proper authorization from Mexico, carry out raids on drug trafficking organizations using American air and ground troops. Colombia clearly overstepped any reasonable lines and this explains Mexico's support for Ecuador on the matter.

As an American, I often feel like my President is unparalleled in stupidity* when it comes to managing our involvement in conflicts and crises around the world. Then, on occasion, someone does something somewhere which makes me realize we are not alone in our experience of being ruled by idiots.

* Ok, that is an overstatement. Bush may be inept at handling many crises but he has made important contributions to the resolution of others.

This week was one such week. Colombia attacked an encampment of FARC personnel in Ecuador during hostage negotiations prompting threats of war from two of Columbia's neighbors, Ecuador and Venezuela. When asked for an appology, Colombia has given given one entirly devoid of substance (sort of "sorry for the inconvenience, but we would do it again").

So the crisis began to unfold. By now at least Ecuador and Venezuela have deployed troops to the border with Colombia, and Peru may be doing the same. The three countries (often at odds with eachother) have united in support for Ecuador's sovereign control over that country's territory. Ecuador has even threatened military action against Colombia if substantive actions are taken on Colombia's side.

This was an extremely stupid move on Colombia's part. This has alienated the country which leases the space for the USAF base which the US uses for most anti-FARC missions, and it has also given Chavez an excuse to provide actual military support to the FARC. Furthermore this isolates Colombia and thus risks to cause issues for the economy of the area. Finally, as long as Ecuador is threatening military action against Colombia over the incursion, this provides FARC with a safe corridor of operations near the border of Ecuador, Peru, and Venezuela. I personally believe that the FARC needs to be defeated but Colombia cannot do this by themselves. I fear this action may cause many issues for Colombia for a long time.

As if this wasn't enough, Ecuador has managed to get all of South America largely on their side. However, Colombia far from being apologetic has decided that they will take the matter to the International Criminal Court where they will charge Chavez with genocide for his alleged (though probable role) in supporting FARC. The problem with this is that the timing at least (and in all likelihood the charges themselves) is so clearly politically motivated that I don't think the ICC would act. Instead this just drives those who are infuriated over the incident into positions where more is required.

I wish I could say that I was optimistic. But I now fear that this will degrade into some sort of war. Colombia has chosen to be the worst kind of neighbor and in all likelihood this will cause serious problems for a long time. If Colombia is to defeat the FARC, they will need all the help they can get. Making enemies with three of their four neighbors does not seem wise.

FARC has scored an important political victory here in the same way that Hezbullah scored an important victory in the war with Israel last year. They have been given, through ineptness on the part of Uribe, a safe haven, and safe corridors of operation. They have also restricted the amount of aid which will be available to Colombia long-term (largely sealing the fate of the lease of the USAF base at Manta, Ecuador), and set their enemies all alone in the night.

For some time I have been looking at the possibility of starting a new organization to help fill what I believe is a gap in the available coverage by other organizations. The Institute for the Advancement of Open Source would:

1) Help with outreach and advocacy of free/open source software, documentation, and content.

2) Provide minimalist and marketing-friendly guidelines and definitions for what qualifies as free/open source. This is different from the OSI OSD in the sense that the OSI OSD is largely designed for evaluating specific licenses by a specific organization. I believe that it is well beyond the time and ability of marketing managers to understand the organization-specific interpretations of that definition (and how the organization's interpretation differs from the nearly identical DFSG).

3) Provide a place where people involved in Free/Open Source software, documentation, and content can come together, work together, and mentor others involved in the same area.

This will be different from other existing organizations in the following ways:

Unlike the FSF, we will have objective criteria for what constitutes Free/Open Source. Nobody will be left wondering how forced advocacy (as in the GNU Manifesto as an invariant section of the EMACS documentation) fits into the free speach/free software world.

Unlike the OSI and SPI, we will not limit ourselves to software. Also OSI has not really taken on a lot of the positive outreach possibilities in recent years, and although SPI has that in its charter, they have not done so either. We will not be providing the organizational support for specific projects that SPI does, nor will we provide license certification as does the OSI. Instead we will be primarily an outreach organization aimed at advancing free/open source sofware.

What do people think? Does this make sense?

holy_calamity writes "Researchers have successfully added two 'unnatural' DNA letters to the code of life. They created two artificial base pairs that are treated as normal by an enzyme that replicates and fixes DNA inside cells. This raises the prospect of engineering life forms with genetic code not possible within nature, allowing new kinds of genetic engineering."

Following disagreements which have gotten me banned from the OSI license discuss list over the right or power of OSI to unilaterally claim total authority over the term "open source" (a view which is disclaimed on the OSI site, but is generally held by various license-discuss participants including the list moderator-- those interested in what was actually said can check the December and January archives of that list on the OSI site), I have decided that the best way forward is to help fill the void in the industry by offering simple and concrete guidelines that my business will be using to determine whether or not we can use a license for development under our open source policy. This policy is generally in the same spirit as OSI's OSD and the FSF's Four Freedoms, but is designed to be less subject to arguments over interpretation, and easier for businesses to use as guidelines for when to state that software is uncontroversially free and open source. Note however that the guidelines are somewhat stricter than either the FSF's guidelines for Free Software or the OSI's OSD.

One of the key points one must be aware of is that software freedom carries with it an economic advantage. The goal of this set of guidelines is to help provide an objective framework for understanding when we feel that this freedom is crippled through onerous requirements either on the developers or the end users. Our commitment is to preserving this freedom for our customers and we hope other businesses will adopt similar guidelines.

The first two requirements are hard requirements. If either of these are violated, we will not do work on the project though we may help arrange others to do this.

1: Open source works must not place restrictions on use, nor may it force one to distribute source code except when one has opted to provide a copy of object code. Furthermore, no bundling restrictions may be in place. This provision disqualifies licenses which restrict commercial use either directly or indirectly (as the Aladdin License seeks to do), as well as ones which force distribution of modifications (such as those of the Aferro GPL and Larry Rosen's Open Software License). It does not disqualify the GPL v2, nor does it fully disqualify the GPL v3.

2: Modifications must be possible for all sections of the work except for the license text itself. One must be allowed to distribute such derivative works and to provide the same rights downstream as were granted to oneself. This disqualifies GFDL works which include invariant sections. If someone stated that one must *choose* a specific license and not provide the rest of the rights downstream, we would not work on that project either. This would also disqualify GPL v3 programs where additional permissions require their own removal on modification.

The following guidelines involve license selection. These do not disqualify us from working on various projects, but help us determine what licenses are best for a project:

3: Licenses should be no more restrictive then absolutely necessary for either party. Where all things are equal, more permissive licenses are preferred.
4: Licenses should be no more complicated than absolutely necessary. Where all things are equal, simpler (and usually shorter) licenses are preferred.

The final guideline defines our community involvement:

5: Multivendor solutions are preferred over single-vendor solutions.

What do people think?

UPDATE:
There has been a change in moderators of the license-discuss list. I was banned by Russ Nelson, not the current moderator (Ernie P). Ernie has been an important positive influence on the OSI in general and I wish him luck. However, I have serious questions about the OSI's ability to actually contribute substantial resources to the community at the present time, so I suppose I will work on these challenges and reconsider my involvement on the lists later.

A ballot initiative in Arazona has been put forward to attempt to deny citizenship to individuals born in this country to parents who are not here legally. I believe that this measure is Unconstitutional and something every American ought to oppose.

To be fair the issue of children being born to illegal immigrants who then give additional benefits to their parents is a real issue, but solutions to this problem exist which do not impose burdens on American citizens and do not run amok with the Constitutional definition of citizenship (being applicable to all who are born or naturalized in this country and who are subject to the jurisdiction of our country) found in the 14th amendment.

Proponents of the measure suggest that individuals born in this country to illegal immigrants do not fall under this ammendment. They point to congressional record to support their arguments but neither the plain wording of the amendment of the 14th Amendment nor the congressional record support this limitation. In general, the 14th amendment was understood to exclude certain classes of individuals, most notabably alians (neither born nor nationalized in the US), and children of diplomats or ministers of foreign governments (whose diplomatic immunity excludes them from US jurisdiction).

An attempt to ensure that only children of US citizens are US citizens would impose serious burdens on US citizens when their children are born. As my son was born overseas, I had to go through a similar process ater his birth. I don't think that requiring something in the US would be wise even if it was Constitutional (which I don't believe it is).

A better approach would be to make a simple rule: US citizens must meet a minimum age of, say, 21 years old before petitioning the entry of blood relatives. Furthermore, when a family of illegal immigrants is detained and deported, any US citizen children of the family who have no immediate relatives legally in the US would be issued a passport and then deported. The passport would then allow the child to apply for a new passport when he/she becomes an adult and re-enter the US as a US citizen. No interpretation of the Constitution I can find would prevent the US government from barring entry of minor US citizens when not accompanied by a legal guardian who is also legally able to enter the country. By placing a minimum age of 21 on top of a petition process which can last several years, we could remove any incentive to have children in the US for the sole purpose of one's own immigration status.

A second piece of our policy needs to be a sane immigration policy which does not create a massive black market for illegal immigration. This black market fuels drug and human trafficking, and these industries could be set up to smuggle other dangerous substances into the US, such as those which might be useful in large-scale terrorist activities. Therefore illegal immigration is a systemic threat to our national security. This means that we need strong reform for the processes of bringing in foreign workers and that we work to slowly help ensure that people who are currently on welfare have an opportunity to make a better living for themselves taking jobs which currently go to illegal immigrants (however, here, the devil is in the details).

NULLs are, IMO, the only major blunder in Codd's initial rules of relational databases. NULLs introduce data ambiguity because they have two possible yet distinct meanings: unknown (but applicable) and not applicable.

Until very recently, I was in the camp which said that use of NULLs should generally be avoided in database design because of the issue raised above (semantic ambiguity). The idea I previously held was that NULLs should only represent information which is presumed to be applicable but is currently known (perhaps because there is no business requirement to collect it). I now believe that a disciplined approach to using NULLs to represent non-applicable values when appropriate is better than a general policy of avoidance. This puts me in a camp now decisively opposite many people in database design (including Chris Date).

The basic issue is that the semantic ambiguity of NULLs makes the relational model incomplete. This has a fairly serious consequence-- many sorts of data constraints cannot be meaningfully applied in relational databases when NULLs cannot be used for non-applicable data. Take for example the following information record:

job_id: 123
created_by: chris
created_at: 2007-12-14 15:30:36.6543-8
completed_at: NULL
success: NULL
error_condition: NULL

we might represent the above record in a table such as:
create table pending_job (
id serial not null unique,
created_by text not null references users(username),
created_at timestamp default now(),
completed_at timestamp,
success bool,
error_condition text,
CHECK (completed_at IS NULL OR success IS NOT NULL),
CHECK (success IS NOT FALSE OR error_condition IS NOT NULL)
);

In short, when the job completes, we must log whether the job was successful.
When the job is not successful, we must log a reason.

Now, it might be possible to break the completed information into a separate table, though I don't see any real normalization benefits to doing so since all dependencies are on the id field anyway (this is artificial data structured in an artificial manner). However, whether or not we do so, we cannot maintain a constraint which states that the error_condition field must be populated on a failed job when it is completed unless we allow the use of NULLs to represent non-applicability in this case.

Cowardly writes "Tasked with managing the ICT requirements of United States President George Bush, an IT professional could be forgiven for allowing their high tech fantasies to go into overdrive. But according to Intercontinental IT manager Ben Wrigley, the job is more like a step back in time.

However, he is quick to point out that reality is very different to all the Hollywood movies featuring the White House as the backdrop to the latest cutting edge technologies used by presidential aides.

In fact, Wrigley said much of the President's communications gear can only be described as "antiquated"."
http://www.computerworld.com.au/index.php/id;1908045667;fp;16;fpid;0

The government of Canada is preparing to attempt to bring a new DMCA-modeled copyright law in Canada in order to comply with the WIPO treaties the country signed in 1997. These treaties were also the base of the American DMCA. The new Canadian law will be even more restrictive in nature than the American version and worse than the last Canadian copyright proposal, the defeated Bill C-60. Amongst the many restrictive clauses, in this new law — as Michael Geist explains — is the total abolishment of the concept of fair use, "No parody exception. No time shifting exception. No device shifting exception. No expanded backup provision. Nothing.". Michael Geist provides a list of 30 things that can be done to address the issue.
http://www.theglobeandmail.com/servlet/story/RTGAM.20071127.WBcyberia20071127170629/WBStory/WBcyberia

Tongues were wagging earlier this week on speculation that the News Corporation would buy LinkedIn, the professionals-oriented social network, early next year.
http://dealbook.blogs.nytimes.com/2007/11/28/how-much-for-linkedin-ceo-says-helluva-lot/index.html?ex=1354078800&en=7caf105be07c7c6f&ei=5088&partner=rssnyt&emc=rss

paleshadows writes "An interesting analysis that has recently went on the air attempts to quantify the popularity of programming langauges through Internet queries. The idea is to query various sources with appropriate strings and use the number of hits as an indicator of popularity. For example, "PHP programmer -"job wanted" site:craigslist.org" was used to estimate the number job postings for PHP programmers. While this methodology is far from being perfect, the results are still interesting. The sources used were Yahoo (estimate general popularity), Craigs List (estimate job offerings), Amazon books (documentation), Freshmeat (popularity within the open-source community), Google Code (also open-source), del.icio.us (social bookmarking), Lamda The Ultimate (academia interest), Programming reddit (programmers talk), and Slashdot. From the nearly 30 langauges surveyed, C turned out to be the most popular in 5 out of the 9 sources, often followed by Java (which is, perhaps surprisingly, the clear #1 in Slashdot). Other top-3 languages are PHP, SQL, VB, C++, JavaScript, Haskell, and Python."
http://langpop.com/

biggknifeparty writes "The NDP government in the Canadian province of Manitoba is introducing legislation requiring anyone who becomes aware of child pornography to report it to the authorities or face up to 2 years in prison and a $50000 dollar fine. Internet Service Providers and computer service shops will now have to forward their suspicions to cybertip.ca. There is no doubt that the intentions are positive, to protect our children above all else, but is this not a slippery slope that will lead to privacy concerns? In Soviet Manitoba citizens spy on each other."

Canada, Denmark, Norway and Russia are among the countries competing for a piece of the Arctic.

http://feeds.wired.com/~r/wired/topheadlines/~3/192214441/st_essay

Having now worked with some interesting applications where artificial data structures and rules were issues (including accounting rules), I have come to decide that RDBMS's need some additional features to meet certain types of needs. Otherwise one is required to forego some (often important) data integrity rules in favor of giving the application authoritative responsibility for data integrity.

The most frequent issue one might have involves double-entry accounting rules. In double-entry accounting, every transaction must be balanced (the sum of the amount must equal 0), but each transaction can involve an arbitrary number of parts. Logically we might have:

Account Number...Amount...Transaction...Reference
1010 Checking...10000....19934.........Total Payment Customer 5
1300 AR.........-3000....19934.........Paying vendor 5, invoice 35
1300 AR.........-7000....19934.........Paying vendor 5, invoice 36

Current generations of RDBMS's that I have worked with do not provide any means of enforcing this sort of constraint outside of procedural logic (i.e. stored procedures). It would be really nice to be able to declare such constraints within the DDL of SQL itself, or at least be able to create triggers which allow this sort of enforcement when running arbitrary SQL.

The major difficulty in enforcing the above transaction is that one can only enforce the constraint at the transaction level. Each individual row insert will violate the constraint. So table-level constraints do not work in this case. I see three options:

1) Deferrable check constraints on the table level. However, performance here is likely to suck because if you have a constraint like:
check sum(amount) = 0 default deferred
every transaction is going to have to run an aggregate on the table, which could become quite large in an accounting applicaion.

2) A better option might be to have an insert_set constraint (when updating is disabled), so that only the inserting record set is checked. Note that this does not affect other situations where these sorts of rules might be in place but updates are allowed.

3) Transaction-level triggers with full access to the old and new row sets. A trigger might look something like this (in pseudocode):

if operation = update then
      select transaction from new where transaction IN (select transaction from affected_table where pkey NOT IN (select pkey from old));
      IF FOUND then
              raise exception "updating part of a transaction not allowed"
      end if;
end if;
select transaction from new group by transaction having sum(amount) 0;
if found then
      raise exception "write set includes unbalanced transaction(s)!"
end if;

While I would prefer a declarative interface (and hence check constraints instead of triggers), it would be nice to have any improvement in this area rather than have to rely on stored procedures to enforce data integrity.