Technical debt is about things like design quality and ease of change. High technical debt code is hard to change. That means that adding new features or fixing bugs is more expensive, takes longer, and is riskier. You pay interest on the debt every time you make a change (or don't make a change because it was too scary because your design sucked).
If you pay off your technical debt by improving the design, adding features and fixing bugs will be cheaper, faster, and less risky.
Atlantis flew a magnificent mission, capping a great career. She, and her sisters, have been great ships and deserve to retire with honour.
Yeah, they were expensive. Yeah, people think robots are cooler. Yeah, they couldn't go to the moon or Mars. And yeah, in hindsight hanging a somewhat fragile spaceship on the side of a booster probably wasn't the best idea.
But Atlantis and her sisters' record of achievement is magnificent, and will probably never be matched. They launched space probes, they conducted research into materials, life sciences, earth sciences, astronomy, and countless other fields. They serviced satellites and space stations, and brought tonnes of equipment back to earth for study and reflight. They provided a convenient platform for experiments and payloads that would otherwise have had to construct their own complete satellites. They did all this 133 times successfully, with only two losses, and in the space business you'd take that success rate any day of the week.
Refactoring code is like paying off debt. When you add new code without thinking about the design, the internal quality goes down. This makes the code harder (read "more expensive", "slower", etc) to work with in the future. And since most of the cost of software is in maintenance rather than the initial development, that's a bigger issue than you think.
Carrying a little bit of technical debt is inevitable, and a good thing. But if you allow too much to build up (by not refactoring debt-heavy areas of code), the interest payments can become crippling. Your code sucks, you can't understand it, and making changes takes far more work than they should. That directly impacts the bottom line.
For non-Aussies, Frontline was a satire of current affairs shows, and the behind the scenes dealing that goes on with them. It was broadcast in the mid-nineties, and was widely acclaimed for being disturbingly true to how things were actually done.
Highly recommended.
The Russian rockets only have similar demonstrated reliability to the shuttle. But still, the shuttle does need to retire. The smart thing to do would be to launch capsules on the EELVs (Atlas 5 or Delta 4), but that has severe political problems (basically, a lot of people would be out of work).
In the meantime, there are essentially a fixed number of shuttle external tanks left. Why not fly those out, whether it takes until 2010 or 2012, whatever, then move on after that?
Burt Rutan makes the observation that when he saw the Redstone rocket at the national air museum he wondered, "why don't we fly this anymore?".
Indeed why! It's cheap, it's simple - simpler can and often does mean safer. The Redstone can get a person or two into orbit.
No, it can't. Redstone could only launch an astronaut on a very short suborbital hop. A substantially larger rocket is needed to get a human into orbit.
Exactly. In the olden days (pre-Win95), a Basic interpreter was right at your fingertips. One was always bundled with a machine, and in the microcomputer days (pre-PC) the main operating environment *was* a Basic interpreter. The barrier to experimentation was effectively zero.
Nowadays, it's much harder to experiment. Sure there's VBScript and the like, but how many people even know they exist? Most of the time, you have to download something. The barrier to entry has gone up steeply, for no good reason.
I really believe that this is going to lead to some nasty problems for our industry in the future. Fewer kids will come into the field. Talent will be harder to find, and more expensive to boot.
Microsoft, do the world a favour. Bundle an accessible (if basic) programming environment with Vista (e.g. Python, Ruby, or even QBasic from DOS 5). Make it visible - stick an icon for it on the Accessories menu - that way curious types will stumble over it and experiment, and hopefully get hooked. Do this, and everyone will forgive you for the Vista delays. I promise;-).
The US modules for the ISS are designed with the shuttle's payload bay in mind. In particular, they expect the loads during launch to be applied from fixtures mounted along the sides and bottom of the bay. For expendable launches, you need to design the modules to take their loads from their bottom end. Converting a module would be very expensive.
If you wanted to, you could launch the module in a sort of adaptor that held the modules as they would be if the shuttle was carrying them. However, that would be heavy, to the point where even a Delta-IV Heavy may have trouble launching the module+adaptor combination.
The two reasons the Russians can get cosmonauts into space for $60m per flight is that wage costs are *much* lower than in the US, and they're flying them on a ship that's vastly less capable than the shuttle. If the US could pay Russian-level wages, the shuttle would be a lot less expensive to fly too.
On the shuttle's safety record, it's in the same ballpark as Soyuz. One accident on Soyuz would tip the balance back in the shuttle's favour. The difference is not significant. Also, Soyuz has had plenty of close calls in recent times.
Dynamic typing basically changes the notion of the type of an object from the question "What are you?" to "What can you do?". As long as the object you're using does something sensible in response to a particular operation, it doesn't matter so much what it is. If it does matter, create the right kind of object in the first place, it's not going to change underneath you.
Practical applications include easier and more flexible mocking, as well as code that is just generally more flexible and simpler to work with. There's often less of it, as well.
But this is a *big* hurtling piece of death. If its orbit takes it too close to anything important, there'll be weeks or months of notice to move other stuff out of the way.
Soyuz *had* the ability to launch 20 times a year. It doesn't now (currently 2 per year). Ramping up the production rate would be costly, and require a lot of new workers, who would have to be trained, etc.
And although Soyuz hasn't killed anyone for a long time, there have been several occasions in the last few years where it came uncomfortably close. Soyuz's record isn't significantly better than the shuttle's.
Of course Slackware is for newbies. I, like probably the majority of people here, started out on Slackware, on a veritable s***box of a PC (486SX/4Mb RAM/120Mb HDD/14.4k modem) downloading floppy sets off the net. That was my first experience with *nix, and damn I liked it!
Slackware has probably done more for getting people into Linux than anything else (including all this media hype of the last few years).
--Chris
Yep, it was the third crew, who spent the best part of three months up there (12 weeks). Early in the flight, one of the crew got spacesick (which happens to 1/2 of all astronauts nowadays, apparently), and they didn't radio the ground to tell them the news. Relations went downhill from there.
Later in the flight, having waaayyy too much work scheduled, the crew went on strike for a day in protest.
They're basically using IBM ThinkPads, running a mixture of custom software (like things to control bolts and radars and stuff, and Caution and Warning alarms) and normal stuff (Word, etc), on a mix of Windows 95 (for bolts and stuff - I'm sure they've had to reboot) and Solaris (for the really important stuff).
The laptops themselves have lotsa hardware mods, see SpaceRef for more info (look under mission guides).
Russia is broke. This is why we are only expecting one more module (DC-1) out of them, and a steady flow of Progress and Soyuz spacecraft, which will basically be funded by the US. At least they're cheap.
Slashdot Mirror
Technical debt is not about bugs, mostly.
Technical debt is about things like design quality and ease of change. High technical debt code is hard to change. That means that adding new features or fixing bugs is more expensive, takes longer, and is riskier. You pay interest on the debt every time you make a change (or don't make a change because it was too scary because your design sucked).
If you pay off your technical debt by improving the design, adding features and fixing bugs will be cheaper, faster, and less risky.
Atlantis flew a magnificent mission, capping a great career. She, and her sisters, have been great ships and deserve to retire with honour.
Yeah, they were expensive. Yeah, people think robots are cooler. Yeah, they couldn't go to the moon or Mars. And yeah, in hindsight hanging a somewhat fragile spaceship on the side of a booster probably wasn't the best idea.
But Atlantis and her sisters' record of achievement is magnificent, and will probably never be matched. They launched space probes, they conducted research into materials, life sciences, earth sciences, astronomy, and countless other fields. They serviced satellites and space stations, and brought tonnes of equipment back to earth for study and reflight. They provided a convenient platform for experiments and payloads that would otherwise have had to construct their own complete satellites. They did all this 133 times successfully, with only two losses, and in the space business you'd take that success rate any day of the week.
Because liquid ones aren't made by a certain company in a certain state. It's all politics.
Had any sense prevailed, we'd be sticking a capsule on top of an existing booster -- Atlas 5 or Delta 4 -- and being done with it.
Not true. not true.
Refactoring code is like paying off debt. When you add new code without thinking about the design, the internal quality goes down. This makes the code harder (read "more expensive", "slower", etc) to work with in the future. And since most of the cost of software is in maintenance rather than the initial development, that's a bigger issue than you think.
Carrying a little bit of technical debt is inevitable, and a good thing. But if you allow too much to build up (by not refactoring debt-heavy areas of code), the interest payments can become crippling. Your code sucks, you can't understand it, and making changes takes far more work than they should. That directly impacts the bottom line.
For non-Aussies, Frontline was a satire of current affairs shows, and the behind the scenes dealing that goes on with them. It was broadcast in the mid-nineties, and was widely acclaimed for being disturbingly true to how things were actually done. Highly recommended.
The Russian rockets only have similar demonstrated reliability to the shuttle. But still, the shuttle does need to retire. The smart thing to do would be to launch capsules on the EELVs (Atlas 5 or Delta 4), but that has severe political problems (basically, a lot of people would be out of work).
In the meantime, there are essentially a fixed number of shuttle external tanks left. Why not fly those out, whether it takes until 2010 or 2012, whatever, then move on after that?
There's one more Hubble repair mission to come (planned for Atlantis's final flight).
And despite future spacecraft like JWST, none of the planned replacements will cover the UV range.
Not even close. Two orders of magnitude lower is a starting point.
On top of that, now Slashdot links to it...
Exactly. In the olden days (pre-Win95), a Basic interpreter was right at your fingertips. One was always bundled with a machine, and in the microcomputer days (pre-PC) the main operating environment *was* a Basic interpreter. The barrier to experimentation was effectively zero.
;-).
Nowadays, it's much harder to experiment. Sure there's VBScript and the like, but how many people even know they exist? Most of the time, you have to download something. The barrier to entry has gone up steeply, for no good reason.
I really believe that this is going to lead to some nasty problems for our industry in the future. Fewer kids will come into the field. Talent will be harder to find, and more expensive to boot.
Microsoft, do the world a favour. Bundle an accessible (if basic) programming environment with Vista (e.g. Python, Ruby, or even QBasic from DOS 5). Make it visible - stick an icon for it on the Accessories menu - that way curious types will stumble over it and experiment, and hopefully get hooked. Do this, and everyone will forgive you for the Vista delays. I promise
The US modules for the ISS are designed with the shuttle's payload bay in mind. In particular, they expect the loads during launch to be applied from fixtures mounted along the sides and bottom of the bay. For expendable launches, you need to design the modules to take their loads from their bottom end. Converting a module would be very expensive.
If you wanted to, you could launch the module in a sort of adaptor that held the modules as they would be if the shuttle was carrying them. However, that would be heavy, to the point where even a Delta-IV Heavy may have trouble launching the module+adaptor combination.
The two reasons the Russians can get cosmonauts into space for $60m per flight is that wage costs are *much* lower than in the US, and they're flying them on a ship that's vastly less capable than the shuttle. If the US could pay Russian-level wages, the shuttle would be a lot less expensive to fly too.
On the shuttle's safety record, it's in the same ballpark as Soyuz. One accident on Soyuz would tip the balance back in the shuttle's favour. The difference is not significant. Also, Soyuz has had plenty of close calls in recent times.
Dynamic typing basically changes the notion of the type of an object from the question "What are you?" to "What can you do?". As long as the object you're using does something sensible in response to a particular operation, it doesn't matter so much what it is. If it does matter, create the right kind of object in the first place, it's not going to change underneath you.
Practical applications include easier and more flexible mocking, as well as code that is just generally more flexible and simpler to work with. There's often less of it, as well.
But this is a *big* hurtling piece of death. If its orbit takes it too close to anything important, there'll be weeks or months of notice to move other stuff out of the way.
It's a non-issue.
Trust me, it's much cheaper just launching new stuff at $10k/kg than to bodge together something using the junk already there.
It's called "space junk" for a reason. Some of it's probably OK, but most of it is real garbage.
I suppose you complain that people were doing "nothing more than a glorified taxi run" when they die on the roads, too?
Face it, most of life is boringly routine, including spaceflight. Not everything has to be about doing bold exploration in to the unknown.
Soyuz *had* the ability to launch 20 times a year. It doesn't now (currently 2 per year). Ramping up the production rate would be costly, and require a lot of new workers, who would have to be trained, etc.
And although Soyuz hasn't killed anyone for a long time, there have been several occasions in the last few years where it came uncomfortably close. Soyuz's record isn't significantly better than the shuttle's.
Actually, the Earth's north magnetic pole is near the southern geographic pole already. We're already "on top".
Of course Slackware is for newbies. I, like probably the majority of people here, started out on Slackware, on a veritable s***box of a PC (486SX/4Mb RAM/120Mb HDD/14.4k modem) downloading floppy sets off the net. That was my first experience with *nix, and damn I liked it! Slackware has probably done more for getting people into Linux than anything else (including all this media hype of the last few years). --Chris
Later in the flight, having waaayyy too much work scheduled, the crew went on strike for a day in protest.
The laptops themselves have lotsa hardware mods, see SpaceRef for more info (look under mission guides).
Russia is broke. This is why we are only expecting one more module (DC-1) out of them, and a steady flow of Progress and Soyuz spacecraft, which will basically be funded by the US. At least they're cheap.