... or at least start out with Firebird, PostgreSQL, or SAP-DB. Even small databases sometimes need real functionality, and these will get you much closer to being able to transition to Oracle or DB2 if you need to. Starting out with MySQL is like starting a shipping company with only a pogo-stick.
My dad recently asked me to help him set up a spreadsheet to track budgets, expenses, etc. for his team. He's pretty dedicated to using linux and openoffice, despite his lack of experience. I tried really, really hard to use the spreadsheet in openoffice to give him what he wanted, but he really needed a database -- a normalized set of tables and some reports, not some end-of-process spreadsheet that can't be re-used for anything. I was almost there, until I discovered the only aggregate I could use inside the spreadsheet, when doing database "stuff" with a spreadsheet as a backend (to keep it simple, make data-entry easy, etc.), was count(). Count() is not terribly useful for financial book-keeping.
He still had MS-Access installed on his windows box, and in a couple hours I had everything set up with a few parameterized queries and crosstab reports, giving him as a -result- what he wanted but giving him a lot more flexibility, none of this "adding a new sheet every month" crap, and so forth. I'm actually rather curious how well some of those Access features work with "real" database back-ends; some of the features offered don't seem to have an obvious SQL equivalent; parameterized queries look like stored procedures to me, but Access would have to know how to create and modify those for each database back-end. Anyway, there's no chance of this database getting big and overflowing Access' limits, and I have to admit liking it.
Quite simply: people use spreadsheets where they shouldn't, they need to be shown they can create "real" databases against which they can run useful queries and yet retain non-redundancy, flexibility, cleanliness. Access does that.
I'll keep using Firebird at work, of course, but there are a lot of things database vendors could learn from Access about language design: letting things be nested easily, parameters that don't suck, at least client-side support for crosstabs, etc. On the other hand, Access is terrible at datatype support, its SQL is limited in ways you might not expect, etc.
All I want is for my users, family, and friends to stop using spreadsheets anytime they want to track something that notepad can't adequately handle. And I just want it to work.
I've been noticing that Wal-Mart's been carrying less and less of the Lego inventory, and what they do carry is Bionicle and Harry Potter, not so much the Designer sets. A customer looking at the meager space allocated to Lego might not even realize that you can still buy good stuff online. Our local toy stores also started carrying less, and now I only buy online for lack of any other useful source. (We actually considered opening our own Lego store!) I don't think Lego entirely did it to itself, I think its retailers also participated. Lego's doing better, they're giving up on franchise themes and moving toward purely Designer sets (bad name, sadly) and buckets of bricks. The sad part is that it might be too late, that retailers and customers will never trust them enough again to carry and buy their products like they once did. Considering the others toys I see for sale at Wal-Mart in particular, I see that as a terrible disservice to kids growing up right now -- they'll wind up with one-use crap that won't teach them anything but violence and greed. (Maybe I'm too harsh, but the classic boys/girls toys really just seem that way to me.)
The problem I've always found with letting users design their own databases (that's when you can even keep them from using Excel to do it) is that they don't understand *basic* cardinality issues. They never think about how many of "these" go to how many of "those" or how they want to avoid re-entering data, make sure they don't modify something they don't mean to, or constrain their numbers to actually line up the way they want them to (assuming they can -- it's amazing how hard it can be to explain that A + B != C in a particular situation, no matter how much you want it to be that way.)
WYSIWYG is fine and good (and even to be encouraged -- databases are not so sacred or complicated that the average joe *shouldn't* be able to build his own to suit his needs,) but could it please come with a good tutorial/wizard that helps the user think through the issues? What entities do you have? How are they related? What changes often? Tutorials and wizards to help them expand the database as their needs change would be good too -- help them move the data into a new layout without breaking anything (while it knows what's going on, it could even create backward-compatible views in some cases), etc. There are a lot of re-used ideas in most databases, related to keeping ordered lists, trees, graphs, non-overlapping date periods, etc. and those should be made easily accessible (and extensible?)
Whatever it is, it needs to be as easy to use as a spreadsheet plus a little bit -- move people away from spreadsheets as much as possible. Having a combined layout + data tool is fun, but it creates absolutely abysmal habits that only result in bad data. And then I have to clean up the mess.
If you read his explanation of plugins in future versions of the filesystem, you'll see he varies between taking the reader for an absolute idiot (explaining things as you would to a three-year-old) and talking about abstract concepts without any context (assuming knowledge,) and then occasionally talking about performance (physical) in the middle of a feature (logical) paragraph (which is detrimental to both endeavours.) There are some good ideas in there, absolutely. He's obviously working to merge file systems with programming languages and databases in terms of extensibility and namespaces, but I think he's still struggling with some of the concepts; I hope he takes his time with it to work it all out, listens to other people, and maybe lets someone else write his documentation for him so he can come across less... badly. It seems pointless to have yet another almost-there-but-not-quite database filesystem. He's so close, I'm sure he can smell it. I think we'll eventually have to do something about file formats themselves; for example, some image types let you store (limited) extra data about the image in the file itself, others don't. When we try to store metadata (data) slightly-outside the file, we wind up duplicating structures in some cases, which leads to redundancy issues. In other cases, no matter how much the authors of a file format thought they were planning ahead, you still can't do what you mean to do with a file's metadata, and you'll have to resort to "wrapping" somehow. This just makes it harder for any sort of search system to know where to look for something. A lot of the issues here really are related to object/relational databases and their evolution -- orthogonality, namespaces, symbols, datatypes, constraints, acid, etc. -- but it seems like researchers in each of the fields are just ignoring each others' work. (As I said, this also relates to programming languages; also consider attempts by microsoft to expose application-specific database internals into the filesystem in WinFS via virtual folders... all interesting stuff.)
It's funny to hear people talk about how users are accustomed to Windows and won't like to switch. Our software is Windows-based, we use linux in the server room. But our users really don't know how to use Windows -- we still wind up teaching them that yes, you can move windows around; you can minimize/maximize them; tab goes between controls; you can drag icons; no, "my documents" is not the only place on your hard drive; no, you shouldn't open any and all files, ever, by first opening Word and then going to "file", "open"... and these are people who have been doing data-entry (on computers) for a decade or more. They don't even catch on to the basics from just sitting there using the operating system for eight hours a day. I think we, as programmers, have lost touch with what it means to get accustomed to something new. We think of it in terms of knowing where everything is in the menus, knowing how files will be laid out after a fresh install, knowing where the configuration panels are, etc. Our users ask us to come and find things in the menus for them, like, say, how to print mailing labels -- something that's in the menu, quite obvious, but they won't see because they refuse to explore. They also refuse to read labels, captions on buttons, or any text longer than three words that you throw at them -- but that's another matter. It's not that they're not capable of learning, they just don't want to. You give them OpenOffice, and they'll use it for six months, and then ask to switch back -- not because they couldn't do anything in particular or because stuff was laid out slightly differently, but just because they don't like the idea of running something other than "real" (Microsoft) Office. Sometimes, I think we should just do the "cold turkey" thing and let them deal with it. I think that's the only way they've made it as far as they have -- at some point they had to move from DOS to Windows, I'm pretty sure they didn't like that either. But they did it, and they usually don't look back by now.
Or we could just expect people not to be stupid: no matter how they're calling it, vote anyway. It's not the media's fault that people are lazy enough to stay home if they think they can't win or have alread won (if we can really call it winning)... publish the data. People only get the government they deserve anyway.
... we could at least reduce it to "the copyright term of the country in which the copyrighted material originated" -- but then someone, somewhere, would decide to make that "one billion years" (assuming the world remembers that copyright has always been a hack, not a natural right, and therefore "forever" is unreasonable) in their country, and everyone else would just publish all of their stuff there, first (a lot like tax havens, etc.)
As I recall, "Sequel" was the name of a product/language that existed back when the SQL standard was being written... someone decided to pick a similar name, to keep continuation (make sure people know that SQL is also about databases)... which is part of why the acronym has meant various things to various people at various times, with no correct answer. As far as I'm concerned, if you say "sequel" you either mean the noun, or you mean the very old product/language, not SQL-92 or any other incarnations of the blasted language.
Also -- why the hell was Firebird left out of the list of "real" databases? None of them are "real" databases anyway -- not datatype-orthogonal, the query language sucks, they mix up physical and logical features all over the place, they don't use candidate-key inference to determine view updatability (noticed how many don't naturally support updates against multi-table views?)... all hacks. Useful hacks, but hacks. At least Firebird and others have real ACID (well, almost -- there are some issues with integrity constraint definitions...)
Indeed. The only accident we ever had was when we tried to re-train our cat to use the human toilet instead of a litter box. He refused. He wound up peeing on things that were his (bed, chair, whatever we never used anyway) and pooped once right in front of the bathroom where his litter box had been up until recently. He won, he got his litter box back, not had a problem since. Our cat and his sister were found on the doorstep of the local vet's office, way too young to be weaned. He's been a little weird emotionally, but potty-training is *not* an issue with cats, even when they've not had a mother caring for them. Now, a genetically-engineered cat who won't steal your post-its of your desk in the middle of the night, that'd be nice. And one who won't wake you up before your alarm clock goes off. And one who doesn't greet you in the morning by biting your foot. And...
I wouldn't consider the Murrah Federal building a place where "few people are". You've not shown a clear delineation of how many people should huddle together for protection before they deserve it, either. Nor would I consider it proof that our security is good to say that we haven't had an attack since 9/11 -- when was the previous one before 9/11? What attacks, precisely, have our counter-measures thwarted? We actually have no clue how good our security is, except for, say, research papers proving that ethnicity-based screening is easy to circumvent and actually counter-productive if your enemies are halfway intelligent (which I know is often doubted.) We've put a lot of people in jail (military or not,) we've given our police forces extra powers we had previously sworn never to give them, but have we actually improved anything?
Nuclear suitcases are exactly why sidewalks are cause for concern -- that's where you'd scare the most people, make them feel vulnerable. This isn't about efficiently killing people -- terrorism is about making you feel fear. (People don't throw their own lives away easily, yet consider Palestinians blowing themselves up and yet killing nobody, or only a few other people. The ratio is terrible, but the effect is palpable.) People react even to relative failures, like France did to the 8 people who died in a series of bombings of the parisian metro in 1995. Police and military in every school and metro station? Sarin gas attacks in the Tokyo metro system killed, what, a dozen people? Was there increased security? Were there further attacks?
Political change seems like a much more effective way of keeping people safe. The Basque ETA will likely try to find new ways to fight for independence, after the Madrid bombings -- the response was such that it was obvious this would no longer be a viable method. The Corsicans, the Northern Irish... even the Algerians seem to be changing their minds about terrorism. Things change, and not just because of physical security -- in fact, I've seen very little evidence (actually none) that increased physical security is really doing anything at all to protect us from, as you put it, "reality."
My worry is that we'll concentrate on one method of protection, ignoring all other possibilities. Our response to 9/11 was to increase security on planes, but we pretty much entirely ignored every other possible vector of attack. We weren't attacked, even though we were wide open. Does that mean anything? The move seemed much more geared to making our population *feel* safe than actually *be* safe -- we had had attacks against planes, but our planes were now safe, so all was well again, obviously. It's an expensive placebo, no? And no, you don't actually feel safe -- if you did, you wouldn't be asking for this protection.
Private bodyguards are already in use; obviously someone thinks they're appropriate and effective. But they're localized, they don't affect those of us who don't want to be affected. And they don't require federal funds. (Yup, even a non-republican can care about how much we spend.)
You're right, terrorists have failed to hurt and scare me enough to react. But if you want to save lives that can be easily saved yet aren't, invest in sending food and medication to countries that need the help, researching new cures, education to prevent spreading diseases and the misuse of guns and drugs, vehicle safety, programs to encourage people to stay healthy (fewer, say, heart attacks), better/cheaper healthcare so people won't go without even when it's available,... (for those who care, I'm not implying which, if any, of the above should or should not be federal programs.)
Yup. Our planes are vulnerable. Our trains are vulnerable. Our buses, our taxies are vulnerable. Any public transportation. I'm not even sure that our sidewalks are safe -- we should remember to track which sidewalks people have been on and scan them for bombs before they can step out onto the street. Mom and pop shops aren't safe either, they should be protected. Every last grocery store,... and every vehicle, for sure. You know how many car bombs get used in, say, the middle-east? Protecting your subway won't prevent that. Oklahoma City bombing? Truck, parking lot, explosives.
The problem is that no matter what you protect, people intent on attacking will just pick another target. Terrorists don't need the targets to be big buildings or subway systems -- they might give up if the only thing left to attack is an abandoned barn in the middle of Kansas, but I'm not even sure of that. Just think of the ways you could go about hurting people, and then think about what good security checkpoints on subways will do. Sure, it'll avoid attacks in the subway (maybe) -- people will feel safe while they're in the subway, then as soon as they come out, they'll panic again. And then we'll have an attack in the subway anyway, and... yeah. You see how it works?
Terrorists win when we're afraid. This shows that we obviously have a problem here -- we're terrified of a few dozen guys somewhere with a chip on their shoulder and some explosives in their backpacks. Is life worth living in constant terror of something that might not happen and that you really can't prevent? You're not safe. We never were safe, we just thought we were.
Now, I think it'd be perfectly appropriate for people who do want the extra security to pool their funds and hire private guards, armored carriers, or whatever it takes to make themselves feel safe. If you want the security, you should at least be able to buy it. You can have your guards check our ID's before letting us anywhere near you. But some of us honestly don't give a damn. It's just not worth the hassle.
Agreed. Copyright extensions should never have been allowed (obvious to most of the crowd here) -- it's a form of copyright violation, effectively asserting a copyright when you have no right to (pretty much on the same scale as me going around deciding to be given copyright over random works in the public domain, or under someone else's control) -- they were only given the copyright in the first place because, as part of the deal, it ended. That's a contract we shouldn't have revised in their favor, it steals from us all.
Seems like with the cross-licensing, there's a chance prices would skyrocket after a few iterations. But then, I don't remember my history well enough to know the consequences of that system... anyone care to elaborate?
2: Watch it multiple times with no additional charge
I think that's the part they're really worried about. You're already (so I hear, haven't seen the legal documentation) expected to only watch a show once if you time-, media-, or location-shift it. However, I do remember reading the ruling about having the right to do the above (shift) and I didn't see any mention of "unless it's provided to you conveniently enough that this doesn't matter" anywhere in the text. Besides, there are other reasons for allowing us to record any media we come across, concerning fair use -- we all have the right to make archival copies of everything, because nothing requires the content producers or distributors to do so; and once it's published, it's destined to go into the public domain eventually -- by archiving, we're protecting our side of the deal (they get exclusive control for N years, we get the content free after that period.)
There are plenty of other fair-use cases, but the content industries are (somewhat understandably) trying to make us forget about them. Then again, it's been argued that if they were to get exactly what they wanted, they'd suddenly find that they couldn't make derivative works of anything (consider how often derivative works have been made from classic stories, particularly by Disney) -- when they run out of original material, they'll be at the mercy of creators and copyright holders forever. Then they'll want public domain again.
Generally I won't see a computer brought to me for virus/spyware/adware removal until it just doesn't work anymore. The last one we saw was running WinME and barely booted -- once booted, it couldn't spare the resources to load drivers for the modem (which was the actual reason for the visit -- no internet connection), nor the cd drives, nor the video card, etc. It was running in 16-color mode, limping along -- that's when it was finally decided it might need something cleaned. Might. No wonder spyware authors aren't worried that users will react violently to stuff... so long as the spyware doesn't actively break anything, it's welcome to install.
At least in europe, shopping carts are tethered to each other (and the racks) by thick chains. To get a cart, you insert a coin -- when I was a kid, it was 10 francs, or about 2 bucks. You get the money back by returning the cart to one of the stands and chaining it back up. Oddly enough, there was no problem with carts all over the parking lot... just tonight, while waiting for a friend outside an american grocery store, we were watching the wind blowing carts around the parking lot, heading for cars... one of my passengers was so kind as to go rescue the poor vehicles by returning the carts to the corral -- two parking spots down. If people can't be so courteous as to return carts to a safe spot -only- two parking spots down, I have a feeling they don't have the smarts to master an onboard computer either. I tend to assume that politeness and courtesy are some of the lowest-level brain functions we have... right above fight/flight.
Haven't we learned from the self-checkout lanes that still require multiple attendants to help with the most basic things? (Like "put in bagging area" actually means "put in bagging area", not "look confused and throw item elsewhere, without pressing the skip button, then don't learn from your mistake when the attendant shows you what you did wrong"?)
I think what you're describing is actually RU486, not the usual kinds of oral contraceptives taken by pretty much every girl I know -- those work on the hormones to prevent the release of the egg from the ovaries in the first place, rather than blocking implantation of a fertilized egg.
I wasn't stating my opinion on the matter, only my reasoning and that of others -- non-religious reasoning. Now you're free to take the debate to its logical ends, including whether or not eggs and sperm should be protected, whether or not sterilization (of any sort) should be allowed, whether or not masturbation is somehow wrong, whether or not a child deserves our protection 9 months, 6 months, 3 months, 5 minutes before birth -- or, heck, we could even say that kids don't even deserve our protection after birth. I know there are a few whose necks I wouldn't mind wringing -- annoying critters. The debate can even include the death penalty, though that one is significantly different in terms of consent (people knowing they're committing a crime for which death is the penalty, doing it anyway, therefore asking for death of their own free will.) Still, it's all connected somehow.
I don't think people are so much worried about the cloning as they are about the research being performed on the embryos. The process of cloning, in nature, aims to produce viable, healthy life. Here, the process is being used for research, with no intention of letting life develop naturally. The problem is this: a human embryo would naturally grow into a human baby/child/adult. We're stopping that so we can dissect it (yes, I realize how few cells are involved). This isn't accidental death (as would happen naturally in a fair number of cases). We protect the children, elderly, and mentally handicapped from anyone who comes knocking with ill intent. Embryos fall under the category of "can't fight for itself", which strikes a cord with anyone who would, under normal circumstances, try to save a child from a predator.
This isn't even a religious thing -- we can see it purely from the standpoint of the state, trying to protect (future?) citizens from harm until the day those citizens can make their own choices (age of consent.) The arguments against this sort of research are the same as against abortion -- we know we're killing something that would otherwise be expected to mature into a consent-capable human. Accidental death is one thing, but intentional death is another.
If you wanted to make it more emotional, you could draw a parallel with some sort of scenario where children are being pulled off the street into dark labs, cut open, and experimented on... and eventually killed, and scientists are complaining that they're not allowed to do valuable research because someone thinks these snot-nosed, wild-eyed, born-to-die-anyway creatures (otherwise known as brats) are to be protected.
This doesn't have to be a religious thing, this isn't about cloning, it does bring back arguments we think we've already resolved (abortion), and no, I don't have the answers. I'm an agnostic, but I still feel this is an important issue -- those who claim it's purely religious should be ashamed of themselves for being so dismissive.
What level of information is appropriate, then? Three candidates is already too many for most people, and I know plenty who'd be just as happy with only one. The fewer the candidates, the less reading to do, the less thinking -- and eventually none. Would four candidates have been acceptable to you? Five? Perhaps rather than dictating to people how many candidates have a chance and therefore matter, we could just provide them with the book/tome and expect them to do their civic duty and read up as much as they can? I'm darned sure they won't, but should we make that decision for them? We shape the minds of electors every time we make this kind of editorial decision, we can't avoid increasing bias so long as we remain in an editorial position. (It's not quite censorship -- this sort of action only blocks communication along one channel, not all.)
As to having a chance of winning -- Dean had a chance of winning for a while, until something was said to the contrary (was it his "temper" that magically threw him out of the realm of acceptability? I don't even remember anymore. That's how quickly his star fell.) Because our voting system forces us to vote out of fear rather than hope, the issue of who is a "viable candidate" is extremely important. Simply saying "he has no chance of winning" is a self-fulfilling prophecy. None of us want to vote for someone without a chance, because we're too afraid of the possible outcomes of doing so.
I'm not even sure this is an accurate estimate of the cost of producing the software from scratch. I downloaded and ran sloccount against a good chunk of our code a few months ago. It reported a cost of around 4 million dollars, 16 programmers, and 2 years. We did it on an average of 3 programmers (with turn-over), in 3 years, for probably around 200k to 300k. We're underpaid (by national averages), but the point is that its version of "how much things normally cost" is at best an average, and more likely just a guesstimate. By the author's own admission (at least last time I read the docs,) sloccount uses cost-estimation algorithms no longer favored by the industry. Even slight changes in coding styles (where you put your braces, as is the case of our C++ code) can have a significant impact on the final estimate. By the method I used, as per the article, the linux kernel would take 176 million dollars to redevelop. Their other estimate was even higher. Why should I trust that estimate more?
What we could do, though, is at least tally the amount of money spent by companies for work done on linux and OSS code -- I know there are several out there who actually have spent hard cash on what is now free, and that's a minimum cost of development based on cold facts.
Interbase/Firebird also supported some other language... GDML, was it? Documentation
As I recall, it's all but impossible to get to anymore. As for multiple languages in the same engine, the SQL support has multiple dialects available, with slightly different support for certain datatypes (particularly date/time-related.) So there's a bit of that already, though not tons.
Business system 12 for an example of yet another language in early systems.
In the case of the owner of my ISP, there's no firing I can do; besides, I like the guy. And yes, web devs as well... all about simple selects and then post-processing everything in the script... after all, it works, right?
In the hacker world, it's not about pretty/good solutions, it's about making a problem go away quickly so you can move on to the next problem. It's all about writing one-time scripts in whatever language you know, not learning another language unless you're bored or can't solve your problem with what you already know... It's not the environment I'm in, but it's one I run into now and then. It's hard to give these people better tools because nothing is "good" to them, but everything's "good enough". Sure, a drill would be perfect for this, but I only have a hammer -- but I've made do a thousand times before, and I can do it again, I don't need a drill. These are people who do, in fact, get problems solved. They're terribly useful/valuable -- they would just be more valuable if you could give them tools with which they could do the job even faster; but you have to get past that initial annoyance and even having to look at a new tool -- a new tool to add to their already long list of unix commands, scripts in random directories, knowledge of several languages, etc. Firing such a person doesn't get the job done.
COBOL was(is) also english-like. People liked it. Then they liked it less. But enough had been written in COBOL that the damn thing just won't die, even if you can do the same thing faster in new languages. Learning a language is trouble, converting old programs is trouble, and you don't want to let people forget old languages for fear nobody will be around to maintain those old, unconverted programs...
SQL will be around for a while still, because it's "good enough" and "already known" and there are lots of "legacy apps using it" and the new stuff is immature. I have trouble convincing people that transactions are important, that joins really should be done in the server, that they shouldn't create attributes named "value1" and "value2" just because they currently only have two,... convincing them that the language they use is (to be nice) underkill is just not going to happen.
Sure, I want a new language, I wouldn't mind learning it, heck, I wouldn't mind writing it. But the problem is convincing enough people that what they have isn't good enough.
Sort of. Some SQL dialects allow for arrays as a datatype, though often without much in the way of good operators. Relational theory allows an attribute (column) to have any domain (type), including complex types like containers (sets, arrays, maps, etc.)... in theory, yes, you should be able to do this. SQL, however, is another story.
Because sometimes good ideas pop up in unexpected places, perhaps? If Hitler himself were to utter words of wisdom (actually good wisdom), would you turn away just to spite him? How very foolish.
You cannot legislate morality: morality, honor, ethics, and law are distinct. Legislation is law: it is a list of punishments for actions. The rest are things you can try to teach people in your home, school, or church. Law is about changing the cost/benefit ratio associated with an action; the rest are about changing your motivations, your conscience, appealing to your wish to "belong", etc. Law may punish that which your morality believes to be wrong, morality/ethics/honor may inform the legislative body, but you cannot legislate morality itself.
Slashdot Mirror
... or at least start out with Firebird, PostgreSQL, or SAP-DB. Even small databases sometimes need real functionality, and these will get you much closer to being able to transition to Oracle or DB2 if you need to. Starting out with MySQL is like starting a shipping company with only a pogo-stick.
My dad recently asked me to help him set up a spreadsheet to track budgets, expenses, etc. for his team. He's pretty dedicated to using linux and openoffice, despite his lack of experience. I tried really, really hard to use the spreadsheet in openoffice to give him what he wanted, but he really needed a database -- a normalized set of tables and some reports, not some end-of-process spreadsheet that can't be re-used for anything. I was almost there, until I discovered the only aggregate I could use inside the spreadsheet, when doing database "stuff" with a spreadsheet as a backend (to keep it simple, make data-entry easy, etc.), was count(). Count() is not terribly useful for financial book-keeping.
He still had MS-Access installed on his windows box, and in a couple hours I had everything set up with a few parameterized queries and crosstab reports, giving him as a -result- what he wanted but giving him a lot more flexibility, none of this "adding a new sheet every month" crap, and so forth. I'm actually rather curious how well some of those Access features work with "real" database back-ends; some of the features offered don't seem to have an obvious SQL equivalent; parameterized queries look like stored procedures to me, but Access would have to know how to create and modify those for each database back-end. Anyway, there's no chance of this database getting big and overflowing Access' limits, and I have to admit liking it.
Quite simply: people use spreadsheets where they shouldn't, they need to be shown they can create "real" databases against which they can run useful queries and yet retain non-redundancy, flexibility, cleanliness. Access does that.
I'll keep using Firebird at work, of course, but there are a lot of things database vendors could learn from Access about language design: letting things be nested easily, parameters that don't suck, at least client-side support for crosstabs, etc. On the other hand, Access is terrible at datatype support, its SQL is limited in ways you might not expect, etc.
All I want is for my users, family, and friends to stop using spreadsheets anytime they want to track something that notepad can't adequately handle. And I just want it to work.
I've been noticing that Wal-Mart's been carrying less and less of the Lego inventory, and what they do carry is Bionicle and Harry Potter, not so much the Designer sets. A customer looking at the meager space allocated to Lego might not even realize that you can still buy good stuff online. Our local toy stores also started carrying less, and now I only buy online for lack of any other useful source. (We actually considered opening our own Lego store!) I don't think Lego entirely did it to itself, I think its retailers also participated. Lego's doing better, they're giving up on franchise themes and moving toward purely Designer sets (bad name, sadly) and buckets of bricks. The sad part is that it might be too late, that retailers and customers will never trust them enough again to carry and buy their products like they once did. Considering the others toys I see for sale at Wal-Mart in particular, I see that as a terrible disservice to kids growing up right now -- they'll wind up with one-use crap that won't teach them anything but violence and greed. (Maybe I'm too harsh, but the classic boys/girls toys really just seem that way to me.)
The problem I've always found with letting users design their own databases (that's when you can even keep them from using Excel to do it) is that they don't understand *basic* cardinality issues. They never think about how many of "these" go to how many of "those" or how they want to avoid re-entering data, make sure they don't modify something they don't mean to, or constrain their numbers to actually line up the way they want them to (assuming they can -- it's amazing how hard it can be to explain that A + B != C in a particular situation, no matter how much you want it to be that way.)
WYSIWYG is fine and good (and even to be encouraged -- databases are not so sacred or complicated that the average joe *shouldn't* be able to build his own to suit his needs,) but could it please come with a good tutorial/wizard that helps the user think through the issues? What entities do you have? How are they related? What changes often? Tutorials and wizards to help them expand the database as their needs change would be good too -- help them move the data into a new layout without breaking anything (while it knows what's going on, it could even create backward-compatible views in some cases), etc. There are a lot of re-used ideas in most databases, related to keeping ordered lists, trees, graphs, non-overlapping date periods, etc. and those should be made easily accessible (and extensible?)
Whatever it is, it needs to be as easy to use as a spreadsheet plus a little bit -- move people away from spreadsheets as much as possible. Having a combined layout + data tool is fun, but it creates absolutely abysmal habits that only result in bad data. And then I have to clean up the mess.
If you read his explanation of plugins in future versions of the filesystem, you'll see he varies between taking the reader for an absolute idiot (explaining things as you would to a three-year-old) and talking about abstract concepts without any context (assuming knowledge,) and then occasionally talking about performance (physical) in the middle of a feature (logical) paragraph (which is detrimental to both endeavours.) There are some good ideas in there, absolutely. He's obviously working to merge file systems with programming languages and databases in terms of extensibility and namespaces, but I think he's still struggling with some of the concepts; I hope he takes his time with it to work it all out, listens to other people, and maybe lets someone else write his documentation for him so he can come across less ... badly. It seems pointless to have yet another almost-there-but-not-quite database filesystem. He's so close, I'm sure he can smell it. ... all interesting stuff.)
I think we'll eventually have to do something about file formats themselves; for example, some image types let you store (limited) extra data about the image in the file itself, others don't. When we try to store metadata (data) slightly-outside the file, we wind up duplicating structures in some cases, which leads to redundancy issues. In other cases, no matter how much the authors of a file format thought they were planning ahead, you still can't do what you mean to do with a file's metadata, and you'll have to resort to "wrapping" somehow. This just makes it harder for any sort of search system to know where to look for something.
A lot of the issues here really are related to object/relational databases and their evolution -- orthogonality, namespaces, symbols, datatypes, constraints, acid, etc. -- but it seems like researchers in each of the fields are just ignoring each others' work. (As I said, this also relates to programming languages; also consider attempts by microsoft to expose application-specific database internals into the filesystem in WinFS via virtual folders
It's funny to hear people talk about how users are accustomed to Windows and won't like to switch. Our software is Windows-based, we use linux in the server room. But our users really don't know how to use Windows -- we still wind up teaching them that yes, you can move windows around; you can minimize/maximize them; tab goes between controls; you can drag icons; no, "my documents" is not the only place on your hard drive; no, you shouldn't open any and all files, ever, by first opening Word and then going to "file", "open" ... and these are people who have been doing data-entry (on computers) for a decade or more. They don't even catch on to the basics from just sitting there using the operating system for eight hours a day. I think we, as programmers, have lost touch with what it means to get accustomed to something new. We think of it in terms of knowing where everything is in the menus, knowing how files will be laid out after a fresh install, knowing where the configuration panels are, etc. Our users ask us to come and find things in the menus for them, like, say, how to print mailing labels -- something that's in the menu, quite obvious, but they won't see because they refuse to explore. They also refuse to read labels, captions on buttons, or any text longer than three words that you throw at them -- but that's another matter. It's not that they're not capable of learning, they just don't want to. You give them OpenOffice, and they'll use it for six months, and then ask to switch back -- not because they couldn't do anything in particular or because stuff was laid out slightly differently, but just because they don't like the idea of running something other than "real" (Microsoft) Office. Sometimes, I think we should just do the "cold turkey" thing and let them deal with it. I think that's the only way they've made it as far as they have -- at some point they had to move from DOS to Windows, I'm pretty sure they didn't like that either. But they did it, and they usually don't look back by now.
Or we could just expect people not to be stupid: no matter how they're calling it, vote anyway. It's not the media's fault that people are lazy enough to stay home if they think they can't win or have alread won (if we can really call it winning) ... publish the data. People only get the government they deserve anyway.
... we could at least reduce it to "the copyright term of the country in which the copyrighted material originated" -- but then someone, somewhere, would decide to make that "one billion years" (assuming the world remembers that copyright has always been a hack, not a natural right, and therefore "forever" is unreasonable) in their country, and everyone else would just publish all of their stuff there, first (a lot like tax havens, etc.)
As I recall, "Sequel" was the name of a product/language that existed back when the SQL standard was being written ... someone decided to pick a similar name, to keep continuation (make sure people know that SQL is also about databases) ... which is part of why the acronym has meant various things to various people at various times, with no correct answer. As far as I'm concerned, if you say "sequel" you either mean the noun, or you mean the very old product/language, not SQL-92 or any other incarnations of the blasted language.
... all hacks. Useful hacks, but hacks. At least Firebird and others have real ACID (well, almost -- there are some issues with integrity constraint definitions ...)
Also -- why the hell was Firebird left out of the list of "real" databases? None of them are "real" databases anyway -- not datatype-orthogonal, the query language sucks, they mix up physical and logical features all over the place, they don't use candidate-key inference to determine view updatability (noticed how many don't naturally support updates against multi-table views?)
Indeed. The only accident we ever had was when we tried to re-train our cat to use the human toilet instead of a litter box. He refused. He wound up peeing on things that were his (bed, chair, whatever we never used anyway) and pooped once right in front of the bathroom where his litter box had been up until recently. He won, he got his litter box back, not had a problem since. Our cat and his sister were found on the doorstep of the local vet's office, way too young to be weaned. He's been a little weird emotionally, but potty-training is *not* an issue with cats, even when they've not had a mother caring for them. Now, a genetically-engineered cat who won't steal your post-its of your desk in the middle of the night, that'd be nice. And one who won't wake you up before your alarm clock goes off. And one who doesn't greet you in the morning by biting your foot. And ...
I wouldn't consider the Murrah Federal building a place where "few people are". You've not shown a clear delineation of how many people should huddle together for protection before they deserve it, either. Nor would I consider it proof that our security is good to say that we haven't had an attack since 9/11 -- when was the previous one before 9/11? What attacks, precisely, have our counter-measures thwarted? We actually have no clue how good our security is, except for, say, research papers proving that ethnicity-based screening is easy to circumvent and actually counter-productive if your enemies are halfway intelligent (which I know is often doubted.) We've put a lot of people in jail (military or not,) we've given our police forces extra powers we had previously sworn never to give them, but have we actually improved anything?
... even the Algerians seem to be changing their minds about terrorism. Things change, and not just because of physical security -- in fact, I've seen very little evidence (actually none) that increased physical security is really doing anything at all to protect us from, as you put it, "reality."
... (for those who care, I'm not implying which, if any, of the above should or should not be federal programs.)
Nuclear suitcases are exactly why sidewalks are cause for concern -- that's where you'd scare the most people, make them feel vulnerable. This isn't about efficiently killing people -- terrorism is about making you feel fear. (People don't throw their own lives away easily, yet consider Palestinians blowing themselves up and yet killing nobody, or only a few other people. The ratio is terrible, but the effect is palpable.) People react even to relative failures, like France did to the 8 people who died in a series of bombings of the parisian metro in 1995. Police and military in every school and metro station? Sarin gas attacks in the Tokyo metro system killed, what, a dozen people? Was there increased security? Were there further attacks?
Political change seems like a much more effective way of keeping people safe. The Basque ETA will likely try to find new ways to fight for independence, after the Madrid bombings -- the response was such that it was obvious this would no longer be a viable method. The Corsicans, the Northern Irish
My worry is that we'll concentrate on one method of protection, ignoring all other possibilities. Our response to 9/11 was to increase security on planes, but we pretty much entirely ignored every other possible vector of attack. We weren't attacked, even though we were wide open. Does that mean anything? The move seemed much more geared to making our population *feel* safe than actually *be* safe -- we had had attacks against planes, but our planes were now safe, so all was well again, obviously. It's an expensive placebo, no? And no, you don't actually feel safe -- if you did, you wouldn't be asking for this protection.
Private bodyguards are already in use; obviously someone thinks they're appropriate and effective. But they're localized, they don't affect those of us who don't want to be affected. And they don't require federal funds. (Yup, even a non-republican can care about how much we spend.)
You're right, terrorists have failed to hurt and scare me enough to react. But if you want to save lives that can be easily saved yet aren't, invest in sending food and medication to countries that need the help, researching new cures, education to prevent spreading diseases and the misuse of guns and drugs, vehicle safety, programs to encourage people to stay healthy (fewer, say, heart attacks), better/cheaper healthcare so people won't go without even when it's available,
Yup. Our planes are vulnerable. Our trains are vulnerable. Our buses, our taxies are vulnerable. Any public transportation. I'm not even sure that our sidewalks are safe -- we should remember to track which sidewalks people have been on and scan them for bombs before they can step out onto the street. Mom and pop shops aren't safe either, they should be protected. Every last grocery store, ... and every vehicle, for sure. You know how many car bombs get used in, say, the middle-east? Protecting your subway won't prevent that. Oklahoma City bombing? Truck, parking lot, explosives.
... yeah. You see how it works?
The problem is that no matter what you protect, people intent on attacking will just pick another target. Terrorists don't need the targets to be big buildings or subway systems -- they might give up if the only thing left to attack is an abandoned barn in the middle of Kansas, but I'm not even sure of that. Just think of the ways you could go about hurting people, and then think about what good security checkpoints on subways will do. Sure, it'll avoid attacks in the subway (maybe) -- people will feel safe while they're in the subway, then as soon as they come out, they'll panic again. And then we'll have an attack in the subway anyway, and
Terrorists win when we're afraid. This shows that we obviously have a problem here -- we're terrified of a few dozen guys somewhere with a chip on their shoulder and some explosives in their backpacks. Is life worth living in constant terror of something that might not happen and that you really can't prevent? You're not safe. We never were safe, we just thought we were.
Now, I think it'd be perfectly appropriate for people who do want the extra security to pool their funds and hire private guards, armored carriers, or whatever it takes to make themselves feel safe. If you want the security, you should at least be able to buy it. You can have your guards check our ID's before letting us anywhere near you. But some of us honestly don't give a damn. It's just not worth the hassle.
Agreed. Copyright extensions should never have been allowed (obvious to most of the crowd here) -- it's a form of copyright violation, effectively asserting a copyright when you have no right to (pretty much on the same scale as me going around deciding to be given copyright over random works in the public domain, or under someone else's control) -- they were only given the copyright in the first place because, as part of the deal, it ended. That's a contract we shouldn't have revised in their favor, it steals from us all.
... anyone care to elaborate?
Seems like with the cross-licensing, there's a chance prices would skyrocket after a few iterations. But then, I don't remember my history well enough to know the consequences of that system
2: Watch it multiple times with no additional charge
I think that's the part they're really worried about. You're already (so I hear, haven't seen the legal documentation) expected to only watch a show once if you time-, media-, or location-shift it. However, I do remember reading the ruling about having the right to do the above (shift) and I didn't see any mention of "unless it's provided to you conveniently enough that this doesn't matter" anywhere in the text. Besides, there are other reasons for allowing us to record any media we come across, concerning fair use -- we all have the right to make archival copies of everything, because nothing requires the content producers or distributors to do so; and once it's published, it's destined to go into the public domain eventually -- by archiving, we're protecting our side of the deal (they get exclusive control for N years, we get the content free after that period.)
There are plenty of other fair-use cases, but the content industries are (somewhat understandably) trying to make us forget about them. Then again, it's been argued that if they were to get exactly what they wanted, they'd suddenly find that they couldn't make derivative works of anything (consider how often derivative works have been made from classic stories, particularly by Disney) -- when they run out of original material, they'll be at the mercy of creators and copyright holders forever. Then they'll want public domain again.
Generally I won't see a computer brought to me for virus/spyware/adware removal until it just doesn't work anymore. The last one we saw was running WinME and barely booted -- once booted, it couldn't spare the resources to load drivers for the modem (which was the actual reason for the visit -- no internet connection), nor the cd drives, nor the video card, etc. It was running in 16-color mode, limping along -- that's when it was finally decided it might need something cleaned. Might. No wonder spyware authors aren't worried that users will react violently to stuff ... so long as the spyware doesn't actively break anything, it's welcome to install.
At least in europe, shopping carts are tethered to each other (and the racks) by thick chains. To get a cart, you insert a coin -- when I was a kid, it was 10 francs, or about 2 bucks. You get the money back by returning the cart to one of the stands and chaining it back up. Oddly enough, there was no problem with carts all over the parking lot ... just tonight, while waiting for a friend outside an american grocery store, we were watching the wind blowing carts around the parking lot, heading for cars ... one of my passengers was so kind as to go rescue the poor vehicles by returning the carts to the corral -- two parking spots down. If people can't be so courteous as to return carts to a safe spot -only- two parking spots down, I have a feeling they don't have the smarts to master an onboard computer either. I tend to assume that politeness and courtesy are some of the lowest-level brain functions we have ... right above fight/flight.
Haven't we learned from the self-checkout lanes that still require multiple attendants to help with the most basic things? (Like "put in bagging area" actually means "put in bagging area", not "look confused and throw item elsewhere, without pressing the skip button, then don't learn from your mistake when the attendant shows you what you did wrong"?)
I think what you're describing is actually RU486, not the usual kinds of oral contraceptives taken by pretty much every girl I know -- those work on the hormones to prevent the release of the egg from the ovaries in the first place, rather than blocking implantation of a fertilized egg.
I wasn't stating my opinion on the matter, only my reasoning and that of others -- non-religious reasoning. Now you're free to take the debate to its logical ends, including whether or not eggs and sperm should be protected, whether or not sterilization (of any sort) should be allowed, whether or not masturbation is somehow wrong, whether or not a child deserves our protection 9 months, 6 months, 3 months, 5 minutes before birth -- or, heck, we could even say that kids don't even deserve our protection after birth. I know there are a few whose necks I wouldn't mind wringing -- annoying critters. The debate can even include the death penalty, though that one is significantly different in terms of consent (people knowing they're committing a crime for which death is the penalty, doing it anyway, therefore asking for death of their own free will.) Still, it's all connected somehow.
I don't think people are so much worried about the cloning as they are about the research being performed on the embryos. The process of cloning, in nature, aims to produce viable, healthy life. Here, the process is being used for research, with no intention of letting life develop naturally. The problem is this: a human embryo would naturally grow into a human baby/child/adult. We're stopping that so we can dissect it (yes, I realize how few cells are involved). This isn't accidental death (as would happen naturally in a fair number of cases). We protect the children, elderly, and mentally handicapped from anyone who comes knocking with ill intent. Embryos fall under the category of "can't fight for itself", which strikes a cord with anyone who would, under normal circumstances, try to save a child from a predator.
... and eventually killed, and scientists are complaining that they're not allowed to do valuable research because someone thinks these snot-nosed, wild-eyed, born-to-die-anyway creatures (otherwise known as brats) are to be protected.
This isn't even a religious thing -- we can see it purely from the standpoint of the state, trying to protect (future?) citizens from harm until the day those citizens can make their own choices (age of consent.) The arguments against this sort of research are the same as against abortion -- we know we're killing something that would otherwise be expected to mature into a consent-capable human. Accidental death is one thing, but intentional death is another.
If you wanted to make it more emotional, you could draw a parallel with some sort of scenario where children are being pulled off the street into dark labs, cut open, and experimented on
This doesn't have to be a religious thing, this isn't about cloning, it does bring back arguments we think we've already resolved (abortion), and no, I don't have the answers. I'm an agnostic, but I still feel this is an important issue -- those who claim it's purely religious should be ashamed of themselves for being so dismissive.
What level of information is appropriate, then? Three candidates is already too many for most people, and I know plenty who'd be just as happy with only one. The fewer the candidates, the less reading to do, the less thinking -- and eventually none. Would four candidates have been acceptable to you? Five? Perhaps rather than dictating to people how many candidates have a chance and therefore matter, we could just provide them with the book/tome and expect them to do their civic duty and read up as much as they can? I'm darned sure they won't, but should we make that decision for them? We shape the minds of electors every time we make this kind of editorial decision, we can't avoid increasing bias so long as we remain in an editorial position. (It's not quite censorship -- this sort of action only blocks communication along one channel, not all.)
As to having a chance of winning -- Dean had a chance of winning for a while, until something was said to the contrary (was it his "temper" that magically threw him out of the realm of acceptability? I don't even remember anymore. That's how quickly his star fell.) Because our voting system forces us to vote out of fear rather than hope, the issue of who is a "viable candidate" is extremely important. Simply saying "he has no chance of winning" is a self-fulfilling prophecy. None of us want to vote for someone without a chance, because we're too afraid of the possible outcomes of doing so.
I'm not even sure this is an accurate estimate of the cost of producing the software from scratch. I downloaded and ran sloccount against a good chunk of our code a few months ago. It reported a cost of around 4 million dollars, 16 programmers, and 2 years. We did it on an average of 3 programmers (with turn-over), in 3 years, for probably around 200k to 300k. We're underpaid (by national averages), but the point is that its version of "how much things normally cost" is at best an average, and more likely just a guesstimate. By the author's own admission (at least last time I read the docs,) sloccount uses cost-estimation algorithms no longer favored by the industry. Even slight changes in coding styles (where you put your braces, as is the case of our C++ code) can have a significant impact on the final estimate. By the method I used, as per the article, the linux kernel would take 176 million dollars to redevelop. Their other estimate was even higher. Why should I trust that estimate more?
What we could do, though, is at least tally the amount of money spent by companies for work done on linux and OSS code -- I know there are several out there who actually have spent hard cash on what is now free, and that's a minimum cost of development based on cold facts.
Interbase/Firebird also supported some other language ... GDML, was it? Documentation
As I recall, it's all but impossible to get to anymore. As for multiple languages in the same engine, the SQL support has multiple dialects available, with slightly different support for certain datatypes (particularly date/time-related.) So there's a bit of that already, though not tons.
Business system 12 for an example of yet another language in early systems.
In the case of the owner of my ISP, there's no firing I can do; besides, I like the guy. And yes, web devs as well ... all about simple selects and then post-processing everything in the script ... after all, it works, right?
... It's not the environment I'm in, but it's one I run into now and then. It's hard to give these people better tools because nothing is "good" to them, but everything's "good enough". Sure, a drill would be perfect for this, but I only have a hammer -- but I've made do a thousand times before, and I can do it again, I don't need a drill. These are people who do, in fact, get problems solved. They're terribly useful/valuable -- they would just be more valuable if you could give them tools with which they could do the job even faster; but you have to get past that initial annoyance and even having to look at a new tool -- a new tool to add to their already long list of unix commands, scripts in random directories, knowledge of several languages, etc. Firing such a person doesn't get the job done.
In the hacker world, it's not about pretty/good solutions, it's about making a problem go away quickly so you can move on to the next problem. It's all about writing one-time scripts in whatever language you know, not learning another language unless you're bored or can't solve your problem with what you already know
COBOL was(is) also english-like. People liked it. Then they liked it less. But enough had been written in COBOL that the damn thing just won't die, even if you can do the same thing faster in new languages. Learning a language is trouble, converting old programs is trouble, and you don't want to let people forget old languages for fear nobody will be around to maintain those old, unconverted programs...
... convincing them that the language they use is (to be nice) underkill is just not going to happen.
SQL will be around for a while still, because it's "good enough" and "already known" and there are lots of "legacy apps using it" and the new stuff is immature. I have trouble convincing people that transactions are important, that joins really should be done in the server, that they shouldn't create attributes named "value1" and "value2" just because they currently only have two,
Sure, I want a new language, I wouldn't mind learning it, heck, I wouldn't mind writing it. But the problem is convincing enough people that what they have isn't good enough.
Sort of. Some SQL dialects allow for arrays as a datatype, though often without much in the way of good operators. Relational theory allows an attribute (column) to have any domain (type), including complex types like containers (sets, arrays, maps, etc.) ... in theory, yes, you should be able to do this. SQL, however, is another story.
Because sometimes good ideas pop up in unexpected places, perhaps? If Hitler himself were to utter words of wisdom (actually good wisdom), would you turn away just to spite him? How very foolish.
You cannot legislate morality: morality, honor, ethics, and law are distinct. Legislation is law: it is a list of punishments for actions. The rest are things you can try to teach people in your home, school, or church. Law is about changing the cost/benefit ratio associated with an action; the rest are about changing your motivations, your conscience, appealing to your wish to "belong", etc. Law may punish that which your morality believes to be wrong, morality/ethics/honor may inform the legislative body, but you cannot legislate morality itself.