A lot of people seem to be saying that kids should be taught Microsoft so they wont need to be retrained when then get jobs. This is inflexible old-think espoused by people who really don't understand how computers work.
I disagree, its a direct result of us (the UK for sure, but I believe other western countries as well) deciding as a matter of public policy that the job of the school system is to train not to educate. This is why you get targets like more vocational qualifications and 50% of people to attend University, as well as teaching Microsoft products, Java, Web design etc in schools and colleges. I happen to believe that this is a huge mistake, but its certainly a popular and prevalent mistake.
Getting it right doesn't mean replacing one brand of software on school computers with another (though blitzing and computer with FrontPage on it would be a start), it means deciding to educate rather than train, to inspire rather than babysit and so on.
I agree entirely about the issues of documenting (and indeed instrumenting) database code (Though I'd characterize forms as a programming language and client side code).
But, but, but
look at the general standard of documentation of traditional client side and server side code - all those reams of clear lucid documentation of how the EJB links to that EJB to this datastore to this web service and so on. The base problem with documentation is that geeks can't write. (or don't understand the app they are building or both).
Oracle code (the RDBMS I work with) can be documented and apps can be documented. That they aren't generally isn't a technology fault but a design and build fault.
complicated doesn't have to imply difficult, though it usually does. driving a car is complicated - but simple enough for hormone fuelled teens with the attention span of a gnat to perform successfully.
Equally most surgery isn't especially complicated - that doesn't make it easy.
I don't often like to plug source access because it's extremely overrated, but as a last resort, if you can instrument your database startup with a debugger and trace the point of failure, you now have an advantage FAR greater than that Oracle is going to give you once while your trouble ticket clears through the dozen support techs who repeat the same useless advice and tie up your time.
I agree with nearly all of your post, though I do think the issue of support is an important one to consider, but I have to point out that You can indeed output extremely detailed trace information about your databases startup, at nearly every level that a debugger will give you. On *nix you could of course truss the process.
Actually if anything there is better reason to be happy about DNA evidence, (so long as lawyers can use statistics responsibly) than there is to be happy about fingerprint evidence. Fingerprint evidence became common currency before courts got very science savvy, so there is in fact little scientific evidence that fingerprints are reliable identifiers. (http://www.newscientist.com/article.ns?id=mg18124 321.100)
Well my objection to the/. 'we want to see the source - its our religion posters' is that they *did* publish the data. The question that arises for me is, so did anyone run it through the competing model and get different results. Here the openness of the code is not really the issue, its the model and especially the predictive power of that model that is important. If the same data gives comparable results then we can conclude that the models are comparable (and then really test them by predicting things). If the same data gives different results then the models are significantly different.
Actually, evidence that came to light into the media at later times showed that Downing Street had, in fact, been told that the information they were basing the attack on was false.
Not at all, they probably knew that it was 'unreliable', not that it was false.
So, the journalist was, in truth, correct. His information and assumptions were correct. Yet Downing Street now expect the BBC to reform because of this political travesty of revealing to the world what was really going on
It was terrible journalism. A single unsubstantiated source apparently made a specific false allegation. The reporter (Gilligan) went live on air with no notes, no corroboration and no evidence and stated that the Downing St press office and not 'intelligence' was the source of the dossier. The BBC deserved to be hung for that. Alistair Campbell did not draw up the dossier.
This is not to defend the war, I don't, but reporting unattributed unsubstantiated tittle tattle on the main opinion forming news program is awful, awful journalism. The reporter was stupid, but the editorial team and the management (who publicly defended their man before they'd even talked to him) were incompetent.
meanwhile there has been no investigation at all into why the Intelligence was so dismal and wrong. The same intelligence services will provide the justifications for house arrest that the govt wishes to introduce. This is the real scandal.
Whilst there often is that correlation, usually IME because they didn't really understand what they were trying to do in the first place, documentation (both comments and the stuff you are supposed to put in docs) is there for more than just helping future coders. It also helps detail what the program is supposed to do and serves as a guide for users and integrators as to the business purpose of the code. comments and docs are good for many reasons.
using 'impossible values' has a couple of specific problems of its own that nulls, or rigorous denormalisation, don't have.
first the impossible happens, as pratchett states 9 times out of 10.
second wrongly skewing data - even worse using the wrong datatype and then skewing the data - really really makes life hard for the query optimizers. at this point database designers usually curse the product for its abysmal performance when they are asking software to understand that 'Dec 31 9999' means 'Unknown', but 'Dec 31 1999' means 'Dead'.
I was wondering about the comment that 'the original designers of the internet had no need to worry about security'. Um wasn't that the military? If they're not concerned about security then god help us.
Actually the atypically high rates of cancers (leukaemia is I think the most studied) come in geographic clumps anyway, that is the normal pattern, which is why it is very difficult to establish cause and effect between the health effects and the power plants.
You can have a look at http://cis.nci.nih.gov/fact/3_58.htm for some of the issues that arise when trying to determine whether something geographically specific is occuring.
Actually the scientist, or at least the mathematician, would probably be interested in the supplemental question given the hardware what is the theoretical minimum time for an algorithm to achieve this output.
This puts bounds on the readers exercise actually finding the damn algorithm.
and of course the war of independence was clothed in fighting for freedom language, but the freedom sought was to replace one form of rule with another homegrown one.
Slashdot Mirror
Well 9i which was released like 3 years ago or something introduced more or less ANSI99 compliant SQL.
You also seem to suggest that SQLServer and DB2 are completely ansi compliant, this is just wrong. No vendor is completely ansi compliant.
Actually Oracle Parallel Server has supported this for years, at least since 8 officially.
I agree entirely about the issues of documenting (and indeed instrumenting) database code (Though I'd characterize forms as a programming language and client side code).
But, but, but
look at the general standard of documentation of traditional client side and server side code - all those reams of clear lucid documentation of how the EJB links to that EJB to this datastore to this web service and so on. The base problem with documentation is that geeks can't write. (or don't understand the app they are building or both).
Oracle code (the RDBMS I work with) can be documented and apps can be documented. That they aren't generally isn't a technology fault but a design and build fault.
complicated doesn't have to imply difficult, though it usually does. driving a car is complicated - but simple enough for hormone fuelled teens with the attention span of a gnat to perform successfully.
Equally most surgery isn't especially complicated - that doesn't make it easy.
I don't often like to plug source access because it's extremely overrated, but as a last resort, if you can instrument your database startup with a debugger and trace the point of failure, you now have an advantage FAR greater than that Oracle is going to give you once while your trouble ticket clears through the dozen support techs who repeat the same useless advice and tie up your time.
I agree with nearly all of your post, though I do think the issue of support is an important one to consider, but I have to point out that You can indeed output extremely detailed trace information about your databases startup, at nearly every level that a debugger will give you. On *nix you could of course truss the process.
In the UK where I am, a Consultant is not an Employee and so would not be subject to Income Tax on that income as an employment.
Actually if anything there is better reason to be happy about DNA evidence, (so long as lawyers can use statistics responsibly) than there is to be happy about fingerprint evidence. Fingerprint evidence became common currency before courts got very science savvy, so there is in fact little scientific evidence that fingerprints are reliable identifiers. (http://www.newscientist.com/article.ns?id=mg18124 321.100)
Well my objection to the /. 'we want to see the source - its our religion posters' is that they *did* publish the data. The question that arises for me is, so did anyone run it through the competing model and get different results. Here the openness of the code is not really the issue, its the model and especially the predictive power of that model that is important. If the same data gives comparable results then we can conclude that the models are comparable (and then really test them by predicting things). If the same data gives different results then the models are significantly different.
so
non-ms products have never suffered from re-introductions of old vulnerabilities.
People who use microsoft are stupider than the average member of the public.
The us and 'old' europe are technological backwaters compared to china (and presumably africa and south america).
and of course all the above comes from an 'insightful' post. well yes i suppose if posturing and random assertions are insightful.
Not at all, they probably knew that it was 'unreliable', not that it was false.
It was terrible journalism. A single unsubstantiated source apparently made a specific false allegation. The reporter (Gilligan) went live on air with no notes, no corroboration and no evidence and stated that the Downing St press office and not 'intelligence' was the source of the dossier. The BBC deserved to be hung for that. Alistair Campbell did not draw up the dossier.
This is not to defend the war, I don't, but reporting unattributed unsubstantiated tittle tattle on the main opinion forming news program is awful, awful journalism. The reporter was stupid, but the editorial team and the management (who publicly defended their man before they'd even talked to him) were incompetent.
meanwhile there has been no investigation at all into why the Intelligence was so dismal and wrong. The same intelligence services will provide the justifications for house arrest that the govt wishes to introduce. This is the real scandal.
Whilst there often is that correlation, usually IME because they didn't really understand what they were trying to do in the first place, documentation (both comments and the stuff you are supposed to put in docs) is there for more than just helping future coders. It also helps detail what the program is supposed to do and serves as a guide for users and integrators as to the business purpose of the code. comments and docs are good for many reasons.
Why do I need three CDs and tweak the operating system just to install the stupid Oracle client?
Because you are
a) using 9i and not 10g
and
b) using the server software installation media and not the client installation media.
using 'impossible values' has a couple of specific problems of its own that nulls, or rigorous denormalisation, don't have.
first the impossible happens, as pratchett states 9 times out of 10.
second wrongly skewing data - even worse using the wrong datatype and then skewing the data - really really makes life hard for the query optimizers. at this point database designers usually curse the product for its abysmal performance when they are asking software to understand that 'Dec 31 9999' means 'Unknown', but 'Dec 31 1999' means 'Dead'.
don't you just love "internet time"
I was wondering about the comment that 'the original designers of the internet had no need to worry about security'. Um wasn't that the military? If they're not concerned about security then god help us.
Actually the atypically high rates of cancers (leukaemia is I think the most studied) come in geographic clumps anyway, that is the normal pattern, which is why it is very difficult to establish cause and effect between the health effects and the power plants.
You can have a look at http://cis.nci.nih.gov/fact/3_58.htm for some of the issues that arise when trying to determine whether something geographically specific is occuring.
Actually the scientist, or at least the mathematician, would probably be interested in the supplemental question given the hardware what is the theoretical minimum time for an algorithm to achieve this output.
This puts bounds on the readers exercise actually finding the damn algorithm.
and of course the war of independence was clothed in fighting for freedom language, but the freedom sought was to replace one form of rule with another homegrown one.