> Once a project is underway, keeping scope in check is critical so you need good project management.
Yes, requirements creep seems to be the main problem. Since software is "intangible", management seems to think that they can change the requirements when it's half done, with no adverse consequences. The same management wouldn't dream of doing the same thing with their new office building. (E.g, doubling the space requirements after the foundation and half the floors have already been built.)
In general, it's a management problem from top to bottom. Start with vague requirements, disallow sufficient time and money even for a minimal implementation of those vague requirements, put underqualified and undisciplined staff on the project, change the vague requirements while the project is in progress, and go on death marches when the inevitable happens.
Software projects aren't going to behave well until management imposes an engineering discipline on them. And the biggest issues for management are (a) deciding what they are going to do before they start, and (b) deciding before they start whether the project is worth the time and money it's going to take, and cancelling it up front if they don't like those times/costs, rather than just trimming the time/cost projections down to something that their organization finds politically acceptable and then going ahead with the project under falsified time/cost projections.
The article says that the impact "must have happened within the past 6,000 years", and then immediately concludes that it is responsible for some specific events 4,300 years ago. Yes, 4300 is "within" the past 6000, but the proposal of cause-and-effect is a rather long stretch until we get the actual date of the crater.
Nor is there anything "mysterious" about the "sudden decline" of the specified nations/dynasties. After all, we know of lots of nations/dynasties that have suddenly declined during the past 6000 years. Do we require meteors to explain them, too?
The basic report of a powerful meteor strike is really interesting -- or at least will be if it is confirmed -- but let's not descend into pseudoscience by "explaining" history with it before there is any evidence to suggest cause-and-effect for specific events.
The claims about Sargonid Akkad seem to be entirely off base anyway. The glory days of Akkad coincided exactly with Sargon's personal reign -- no rare occurence in ancient history. Moreover Akkad saw a revival just a few decades later, during the reign of his grandson Narim-Sin. Not long after that Akkad did collapse altogether, but that can be explained by the ravages of Guti highlanders, without having to invoke meteors, divine wrath, aliens, or Microsoft's predatory marketing.
People are too quick to invoke grand catastrophes to "explain" things that don't need explaining in the first place. Let's stay skeptical until there is some actual evidence for something.
Also, notice that the article was dated back in April. Any more recent publications on it, anyone?
> A free software project can't show a business need... This is an effective lock-out for free software for all of the subsystems mentioned.
So long as free software is free, it won't generate much money for campaign contributions. So long as it doesn't generate much money for campaign contributions, its well being will be completely ignored by government institutions.
> Why should someone pay lots of money for your product that implements an amazing new encoding algorithm that you payed a million dollars to develop when they could use Free Software that does the same thing
How many algorithms do you know of that someone spent a million dollars developing?
Most of the really important algorithms, just like mathematics, are coming out of public research institutions (aka "universities") and are published without patent encumbrance.
Unfortunately the PTO has succumbed to pressure from money interests and made lots of formerly unpatentable stuff patentable, opening up a gold rush.
No, not an innovation rush -- a gold rush.
Also, notice that (in the USA at least) patents are justified as a way to promote commerce. Commerce, like innovation, was absolutely booming in IT even without algorithm patents.
As Alan Cox said, patents are just a way for the big guys to keep the little guys out. The big guys all hold lots of patents, so they can pay each other off with funny money (or by swapping licenses for lawsuits), but the little guy has to pay real money if he wants to play.
The rich get richer by natural processes; patents are a government institution that serves primarily to speed the process up.
Just like with the DoJ's pact with Microsoft, the US government is more interested in floating the stock market than in supporting economic fair play, or even long-term economic well being.
> that MS gets months every time they have to prepare a statement but the states are now asked to make their most important decision in just a couple of days?
Sounds like we got stuck with a judge that thinks "resolution" is more important that "solution".
Microsoft wins again. Stay tuned for the next round in five years... assuming there's any competition left for the states to care about in five years.
> Switching from Windows to *nix is a 100% change in platform. That's VERY VERY expensive to do, which is why you'll find almost nobody doing it.
Yeah, and companies never have to retrain their staffs or rework their software when they upgrade Windows, do they?
As usual, the people you describe are pinching pennies for the short run and costing themselves dearly over the long run, by sticking with a system owned by a robber baron and changed at his whims, rather than moving to a system based on open standards.
> In the point of view of MS, anyone who's not using a MS product is concidered a failure.
Yep. The biggest problem with MSFT for the past few years is that MS has saturated the desktop and stalled out in its grab for server space. MS needs growth to keep those inflated share prices up.
Also, even though Linux is "competing" mostly with Sun these days, every time an organization adopts Linux for any reason Linux becomes more visible and more credible.
Those who wish to view the history of the universe as a struggle between Linux and Microsoft can still see this as Linux moving to consolidate its base by rallying the rebels and independents throughout {the galaxy, Middle Earth} before going over to the offensive against the strongholds of the {Empire, Dark Lord}.
U.S. District Judge Colleen Kollar-Kotelly, the new trial judge, had set a deadline of Friday for any settlement, citing ``the recent tragic events affecting our nation.''
Once more, the "national tragedy" is invoked as an excuse to give the powerful what they want.
> But you're missing my question. How do you distinguish between child elements (Name) and attributes (Type, worthless-stock-options).
I'll acknowledge that attributes may have some use that hasn't occured to me which is driving your question, but as it is I have never seen any reason to split out "attributes" and "children" into separate semantic classes. Attributes look to me like extra syntax thrown at a non-existent problem.
If I really needed to make that distinction, I'd just do:
would work. In fact that's what I would probably do (depending on exactly what I needed to represent).
Notice that if you have already found the Editor structure, you can take the cdr to get a list of key-value pairs, and use assoc to find the key-value pair that you want.
This can be abstracted pretty easily into a hierarchy of lookup tables, and you can write really simple functions to extract the parts you want.
Yes, and I can do the same thing in Scheme with about half as many characters, and with the added advantage of being able to treat parts my data and stylesheet as executable code if I wish.
Nor do I have to reformat it with a bunch of ampersands to post it to Slashdot, by the way.
(SlashDot
(Editor
(Name CmdrTaco)
)
)
Even more readable, IMO.
Oh, and people had been doing this for years before XML was ever misbegotten.
Slashdot Mirror
> started writing a paper about this topic once, but I never finished it.
Me t
> Once a project is underway, keeping scope in check is critical so you need good project management.
Yes, requirements creep seems to be the main problem. Since software is "intangible", management seems to think that they can change the requirements when it's half done, with no adverse consequences. The same management wouldn't dream of doing the same thing with their new office building. (E.g, doubling the space requirements after the foundation and half the floors have already been built.)
In general, it's a management problem from top to bottom. Start with vague requirements, disallow sufficient time and money even for a minimal implementation of those vague requirements, put underqualified and undisciplined staff on the project, change the vague requirements while the project is in progress, and go on death marches when the inevitable happens.
Software projects aren't going to behave well until management imposes an engineering discipline on them. And the biggest issues for management are (a) deciding what they are going to do before they start, and (b) deciding before they start whether the project is worth the time and money it's going to take, and cancelling it up front if they don't like those times/costs, rather than just trimming the time/cost projections down to something that their organization finds politically acceptable and then going ahead with the project under falsified time/cost projections.
> Gomorrah. That's why sodomy still exists - I don't even want to think what Gomorramy was.
Actually, gomorrahmy simply went out of style because people didn't enjoy it as much as they enjoyed sodomy.
The article says that the impact "must have happened within the past 6,000 years", and then immediately concludes that it is responsible for some specific events 4,300 years ago. Yes, 4300 is "within" the past 6000, but the proposal of cause-and-effect is a rather long stretch until we get the actual date of the crater.
Nor is there anything "mysterious" about the "sudden decline" of the specified nations/dynasties. After all, we know of lots of nations/dynasties that have suddenly declined during the past 6000 years. Do we require meteors to explain them, too?
The basic report of a powerful meteor strike is really interesting -- or at least will be if it is confirmed -- but let's not descend into pseudoscience by "explaining" history with it before there is any evidence to suggest cause-and-effect for specific events.
The claims about Sargonid Akkad seem to be entirely off base anyway. The glory days of Akkad coincided exactly with Sargon's personal reign -- no rare occurence in ancient history. Moreover Akkad saw a revival just a few decades later, during the reign of his grandson Narim-Sin. Not long after that Akkad did collapse altogether, but that can be explained by the ravages of Guti highlanders, without having to invoke meteors, divine wrath, aliens, or Microsoft's predatory marketing.
People are too quick to invoke grand catastrophes to "explain" things that don't need explaining in the first place. Let's stay skeptical until there is some actual evidence for something.
Also, notice that the article was dated back in April. Any more recent publications on it, anyone?
> A free software project can't show a business need... This is an effective lock-out for free software for all of the subsystems mentioned.
So long as free software is free, it won't generate much money for campaign contributions. So long as it doesn't generate much money for campaign contributions, its well being will be completely ignored by government institutions.
> Why should someone pay lots of money for your product that implements an amazing new encoding algorithm that you payed a million dollars to develop when they could use Free Software that does the same thing
How many algorithms do you know of that someone spent a million dollars developing?
Most of the really important algorithms, just like mathematics, are coming out of public research institutions (aka "universities") and are published without patent encumbrance.
Unfortunately the PTO has succumbed to pressure from money interests and made lots of formerly unpatentable stuff patentable, opening up a gold rush.
No, not an innovation rush -- a gold rush.
Also, notice that (in the USA at least) patents are justified as a way to promote commerce. Commerce, like innovation, was absolutely booming in IT even without algorithm patents.
As Alan Cox said, patents are just a way for the big guys to keep the little guys out. The big guys all hold lots of patents, so they can pay each other off with funny money (or by swapping licenses for lawsuits), but the little guy has to pay real money if he wants to play.
The rich get richer by natural processes; patents are a government institution that serves primarily to speed the process up.
Just like with the DoJ's pact with Microsoft, the US government is more interested in floating the stock market than in supporting economic fair play, or even long-term economic well being.
> that MS gets months every time they have to prepare a statement but the states are now asked to make their most important decision in just a couple of days?
Sounds like we got stuck with a judge that thinks "resolution" is more important that "solution".
Microsoft wins again. Stay tuned for the next round in five years
> Switching from Windows to *nix is a 100% change in platform. That's VERY VERY expensive to do, which is why you'll find almost nobody doing it.
Yeah, and companies never have to retrain their staffs or rework their software when they upgrade Windows, do they?
As usual, the people you describe are pinching pennies for the short run and costing themselves dearly over the long run, by sticking with a system owned by a robber baron and changed at his whims, rather than moving to a system based on open standards.
> Where's my credit?? ;-)
I'd mod you up, but I already posted to this thread to say you were off topic.
> In the point of view of MS, anyone who's not using a MS product is concidered a failure.
Yep. The biggest problem with MSFT for the past few years is that MS has saturated the desktop and stalled out in its grab for server space. MS needs growth to keep those inflated share prices up.
Also, even though Linux is "competing" mostly with Sun these days, every time an organization adopts Linux for any reason Linux becomes more visible and more credible.
Those who wish to view the history of the universe as a struggle between Linux and Microsoft can still see this as Linux moving to consolidate its base by rallying the rebels and independents throughout {the galaxy, Middle Earth} before going over to the offensive against the strongholds of the {Empire, Dark Lord}.
> Watch concerns over "the economy" and vague "terrorist" tie-ins put the brakes on the states actions.
Too late:Once more, the "national tragedy" is invoked as an excuse to give the powerful what they want.
> They replaced Solaris boxen with Linux boxen. This, actually, has nothing to do with Microsoft.
Yes, but we notice that they didn't move to Microsoft.
Where's Microsoft's master plan headded if people move from big pricey UNIX to little cheap Linux, instead of to medium medium Windows?
This is just another sign that Microsoft's attempt to 0wn server space is stalling out.
> But I think the thing that gets overlooked the most is how good the IT staff at any particular company/institution might be.
If your IT staff sux0rs, you're going to get f0rked regardless of what OS you run.
> Its hard to imagine a fast-moving cloud of fine dust particles causing such damage.
Never been on the wrong end of a sandblasting machine, have you?
> But you're missing my question. How do you distinguish between child elements (Name) and attributes (Type, worthless-stock-options).
I'll acknowledge that attributes may have some use that hasn't occured to me which is driving your question, but as it is I have never seen any reason to split out "attributes" and "children" into separate semantic classes. Attributes look to me like extra syntax thrown at a non-existent problem.
If I really needed to make that distinction, I'd just do:
(blah
(attributes (a1 blah) (a2 blah)
(child1 blah)
(child2 blah)
)
or something like that.
> Well that breaks stuff, right? If elements with attributes start with two parentheses that makes them different from elements without attributes.
I wouldn't use the double parens. Something like -
(Editor
(type Full-Time)
(worthless-stock-options yes)
(name CmdrTaco)
)
would work. In fact that's what I would probably do (depending on exactly what I needed to represent).
Notice that if you have already found the Editor structure, you can take the cdr to get a list of key-value pairs, and use assoc to find the key-value pair that you want.
This can be abstracted pretty easily into a hierarchy of lookup tables, and you can write really simple functions to extract the parts you want.
> Is "Attack of the Clones" any less cheesy than "Empire Strikes Back"? Or "The Phantom Menace"?
No, it's more cheesy.
> I had much the same ominous feeling that I had at the end of Empire Strikes Back...you don't know where the "good guys" are going from here.
Yeah... <dreamyvoice>back when we still thought Lucas was one of the good guys...</dreamyvoice>
> Well, it isn't really news, but this is a clear sign of what audience the movie is targeted at.
Actually, I have Monsters, Inc. on my calendar.
Hope the SW trailer doesn't ruin it for me.
> And the result is lousy.
See my response to the thread "Close minded people sadden me."
> But with xml, you can store this bit of information with context:
<SlashDot>
<Editor>
<Name>CmdrTaco</Name>
</Editor>
</Slashdot>
Isn't that more informative?
Yes, and I can do the same thing in Scheme with about half as many characters, and with the added advantage of being able to treat parts my data and stylesheet as executable code if I wish.
Nor do I have to reformat it with a bunch of ampersands to post it to Slashdot, by the way.
(SlashDot
(Editor
(Name CmdrTaco)
)
)
Even more readable, IMO.
Oh, and people had been doing this for years before XML was ever misbegotten.
XML: More snake oil to the rescue.
> > slashdot sometimes runs duplicates to see who's still awake on a weekend.
> I thought this was for the benefit of those people who only read Slash Dot on the weekends.
I thought it was for the benefit of editors who don't read Slashdot at all!
> Looks like the same [slashdot.org] monopoly we read about just a couple of days ago on
But how can it be a monopoly if there are two of them!
> I can't believe the marketing idio..er..moro..um..people at Microsoft didn't consider it.
Don't you realize that 'marketing' was the offensive word in that sentence?
ps - I coulda made it funnier, but it mighta given offense.
Yep. All that's left is the "eternal vigilance" part.
Which is where Americans usually drop the ball.