I have to pretty much reverse engineer it from the source code to figure out what you *actually* wrote vs. what you were supposed to write--then I usually end up re-writing it from the original requirements to do what it should have been doing in the first place, because the existing code is such a mess.
In my experience, by the time the code has been through acceptance testing, the behaviour of the messy existing code is a far better indication of what is needed than the requirements documents.
Documentation is rarely subject to the same level of quality control as delivered code. It's more prone to being vague, incomplete or out-of-date.
I think the way forward is to replace detailed documentation with a working prototype, written in an expressive high-level language.
Changing the DPI was one of the first things I tried on Vista, and found it badly broken.
When the mouse pointer was outside a window, the pointer location was reported to the window incorrectly. This meant that when dragging a scrollbar, you had to be careful not to stray outside the window, or the scrollbar would jump.
Hopefully this is fixed by now.
A similar disappointment was that the screen magnifier still showed everything pixellated. I had expected it to re-render a lot of stuff at the new resolution.
In all seriousness, this is basically what I'm doing - but only for a moderate amount of data (< 5GB) that's specifically under my control.
Pros: - searching is fast and centralized (i.e. one search covers the lot) - content is accessible at local disk speeds - network or file server outages aren't a big deal - 'local' files are included in file server backups
The software I'm using is:
sync: SyncBackSE search: Copernic Desktop Search Professional
My resyncs are scheduled nightly. I'd prefer to have local changes syncing live to the network shares (and not vice-versa), but haven't found sync software that does this efficiently on Windows.
The only major downside is the mess when I have to use the network shares directly, before my local changes are synced across. To work around this I'll sometimes edit the remote copy directly, or kick off a resync.
A big problem with that idea is that banks aren't going to take your diploma as security. An unsecured loan isn't going to get the low interest rates available for home loans.
The best you can do is tap a home loan to pay off your student debt, as someone already suggested.
It sounds to me like there's a lack of competition in the US student loan business. Is there anything stopping private interests from funding people through college in exchange for a cut of future earnings, and undercutting the government offering?
I've long regarded the primary purpose of state-run schooling to be to get young people out of the way so that their parents can get on with contributing to society.
Providing education services is a lesser concern, and it shows.
After learning the basics of writing Chinese characters both by hand and by computer, I decided that becoming proficient at handwriting was just not worth it. Sure, it's a very useful skill, but it takes a lot of learning (over a thousand hours) and ultimately doesn't give much benefit.
With the aid of a computer, you don't need to remember all the components and stroke orders for each character. You just need to know how to pronounce what you want to write, and be able to distinguish between different characters with the same pronunciation at sight. If you can both speak and read, you get the harder skill of writing for free.
I use my study time for reading instead of writing.
The same argument can be made against becoming a really proficient speller in English. Really you only need the basics, and be able to deal well with homonyms. Your computer will get you the rest of the way to near-perfect spelling.
My tot is going on two years, and I've decided that as soon as he's old enough to be out alone without adult supervision, I'm more than happy for him to carry a phone.
He mightn't be old enough to use SMS, but I'll make sure he knows how to call me on it, and also make emergency calls.
This isn't about being a control-freak parent, in fact the opposite. This is about giving the young one a degree of freedom to discover the world on his own terms, while keeping a safety net in place in case anything goes wrong.
Anonymous inner classes are clearly a hack necessitated by a lack of first-class function types. Lack of a more concise capturing mechanism (i.e. true lambdas) is crippling.
I can't agree that anonymous functions make more sense than anonymous classes in Java. Java is an OO language and code consists almost entirely of class declarations. Anonymous classes are a good fit for this paradigm, just as lambda expressions are great in functional languages.
I agree entirely that closures and higher-order functions would be an improvement for cases where only a single method is needed (e.g. Runnable), but that's just a special case. Anonymous classes are more general.
If large swathes of your code are auto-generated, that's a clear sign that you're coding at the wrong level of abstraction. And unfortunately Java isn't powerful enough to allow code that is both idiomatic and terse.
Don't forget that most of the development workload is in maintenance, for which all the generated code *does* add significant overheads.
BTW, have you noticed that your auto-generated JavaDocs suck?
Kid: Try to listen and learn. Check that ego. Come off it, I'm the prophet, the professor, I'ma teach you about the worm, who eventually turned to catch wreck with the neck of a long-time oppressor. And he's running from the devil, but the debt is always gaining, and if he's worth being hurt, he's worth bringin' pain in. When the sunshine don't work, the good Lord bring the rain in. Cop: Ok, whatever that meant. I'm sure that's real helpful, Ice-T.
Raytracing scales O(pixels) and O(ln(complexity)). Rasterization is relatively constant in the number of pixels, and O(complexity).
That being the case, and contrary to what many are claiming, ray tracing looks to be the loser in terms of computational complexity...
If we assume scene complexity increases at most in proportion to pixel count (seems reasonable, as we don't want to waste effort describing sub-pixel details), we get:
Agreed, being able to type a bit faster is nice, but the main benefit of Dvorak for heavy usage is the improved ergonomics.
Re-training in Dvorak might not be cost-effective for the employer, but how about the poor employee (and their health provider) who gets their wrists f*cked from typing QWERTY all day?
I'm not going to advocate Dvorak for new typists though, because much like QWERTY it was designed for manual typewriters, and computer keyboards are different.
Dvorak tends to split digraphs across the two hands. This is good for mechanical typewriters that require force from the wrist. For computer keyboards, it's faster to type pairs of adjacent letters on the same hand.
Maybe something like 'Snaither' here...
1 2 3 4 5 6 7 8 9 0 [ ] L F O P J Z M D Y W / = S N A I . , T H E R - G U Q B ' ; V K X C
What I wanna do is set up a honey pot dresser drawer that's wired to a silent alarm, maybe one that sends me an SMS and activates surveillance cameras.
The contents of the honey pot drawer would of course include something of value with a homing beacon concealed in it.
(But maybe I won't bother now everyone knows how secure my house is.)
Slashdot Mirror
I have to pretty much reverse engineer it from the source code to figure out what you *actually* wrote vs. what you were supposed to write--then I usually end up re-writing it from the original requirements to do what it should have been doing in the first place, because the existing code is such a mess.
In my experience, by the time the code has been through acceptance testing, the behaviour of the messy existing code is a far better indication of what is needed than the requirements documents.
Documentation is rarely subject to the same level of quality control as delivered code. It's more prone to being vague, incomplete or out-of-date.
I think the way forward is to replace detailed documentation with a working prototype, written in an expressive high-level language.
I like that site a lot more since I realized what extreme lengths they go to in order to stop me viewing the solution comments without registering.
Getting it to work perfectly across all major browsers (e.g. Firefox) must have been a major technical challenge.
Changing the DPI was one of the first things I tried on Vista, and found it badly broken.
When the mouse pointer was outside a window, the pointer location was reported to the window incorrectly. This meant that when dragging a scrollbar, you had to be careful not to stray outside the window, or the scrollbar would jump.
Hopefully this is fixed by now.
A similar disappointment was that the screen magnifier still showed everything pixellated. I had expected it to re-render a lot of stuff at the new resolution.
I'll wager this isn't going to happen. An Oracle-branded MySQL would be a baffling addition to their product line-up.
This wouldn't get past the marketing guys - MySQL is pretty much everything that Oracle proudly isn't.
Maybe the plan is to reassign the MySQL talent to work on the Oracle database.
In all seriousness, this is basically what I'm doing - but only for a moderate amount of data (< 5GB) that's specifically under my control.
Pros:
- searching is fast and centralized (i.e. one search covers the lot)
- content is accessible at local disk speeds
- network or file server outages aren't a big deal
- 'local' files are included in file server backups
The software I'm using is:
sync: SyncBackSE
search: Copernic Desktop Search Professional
My resyncs are scheduled nightly. I'd prefer to have local changes syncing live to the network shares (and not vice-versa), but haven't found sync software that does this efficiently on Windows.
The only major downside is the mess when I have to use the network shares directly, before my local changes are synced across. To work around this I'll sometimes edit the remote copy directly, or kick off a resync.
A big problem with that idea is that banks aren't going to take your diploma as security. An unsecured loan isn't going to get the low interest rates available for home loans.
The best you can do is tap a home loan to pay off your student debt, as someone already suggested.
It sounds to me like there's a lack of competition in the US student loan business. Is there anything stopping private interests from funding people through college in exchange for a cut of future earnings, and undercutting the government offering?
I've long regarded the primary purpose of state-run schooling to be to get young people out of the way so that their parents can get on with contributing to society.
Providing education services is a lesser concern, and it shows.
After learning the basics of writing Chinese characters both by hand and by computer, I decided that becoming proficient at handwriting was just not worth it. Sure, it's a very useful skill, but it takes a lot of learning (over a thousand hours) and ultimately doesn't give much benefit.
With the aid of a computer, you don't need to remember all the components and stroke orders for each character. You just need to know how to pronounce what you want to write, and be able to distinguish between different characters with the same pronunciation at sight. If you can both speak and read, you get the harder skill of writing for free.
I use my study time for reading instead of writing.
The same argument can be made against becoming a really proficient speller in English. Really you only need the basics, and be able to deal well with homonyms. Your computer will get you the rest of the way to near-perfect spelling.
My tot is going on two years, and I've decided that as soon as he's old enough to be out alone without adult supervision, I'm more than happy for him to carry a phone.
He mightn't be old enough to use SMS, but I'll make sure he knows how to call me on it, and also make emergency calls.
This isn't about being a control-freak parent, in fact the opposite. This is about giving the young one a degree of freedom to discover the world on his own terms, while keeping a safety net in place in case anything goes wrong.
Anonymous inner classes are clearly a hack necessitated by a lack of first-class function types. Lack of a more concise capturing mechanism (i.e. true lambdas) is crippling.
I can't agree that anonymous functions make more sense than anonymous classes in Java. Java is an OO language and code consists almost entirely of class declarations. Anonymous classes are a good fit for this paradigm, just as lambda expressions are great in functional languages.
I agree entirely that closures and higher-order functions would be an improvement for cases where only a single method is needed (e.g. Runnable), but that's just a special case. Anonymous classes are more general.
Mod parent up.
The overall costs of having to physically go to work each day are vast, greatly outweighing any downsides of telecommuting.
Think what it would mean for the whole property boom/bust mess if people could live anywhere they liked...
If large swathes of your code are auto-generated, that's a clear sign that you're coding at the wrong level of abstraction. And unfortunately Java isn't powerful enough to allow code that is both idiomatic and terse.
Don't forget that most of the development workload is in maintenance, for which all the generated code *does* add significant overheads.
BTW, have you noticed that your auto-generated JavaDocs suck?
So you don't care about how fast trains are because it's faster to fly.
But wait, if trains suddenly became ten faster, you'd consider taking the train, right? So you do care.
I'm sorry, but how is that informative?
'Bollocks', 'rubbish' and 'arse' are also British slang, and 'crap' is ambiguous without the others.
Try 'hogwash'.
Stink as, bro!
Kid: Try to listen and learn. Check that ego. Come off it, I'm the prophet, the professor, I'ma teach you about the worm, who eventually turned to catch wreck with the neck of a long-time oppressor. And he's running from the devil, but the debt is always gaining, and if he's worth being hurt, he's worth bringin' pain in. When the sunshine don't work, the good Lord bring the rain in.
Cop: Ok, whatever that meant. I'm sure that's real helpful, Ice-T.
The unwanted child cries unheard, it's life ends too soon.
Actually in mandarin the bread 'bao' is first tone (high-pitch, level inflection).
In normal speech, the tones are pretty well unambiguous to native speakers, so the scope for punning and confusion is less than one might think.
Aside for normal speech, there are situations where tones are ambiguous:
- singing - no tones, although I wonder if there might be some statistical relationship between tones and melody contours
- whispering - tones only partially discernible
- 'pinyin' phonetic spelling using the latin alphabet, if tone-marks aren't used
That last one is a pet peeve of mine. The use of pinyin tone-marks is uncommon outside of text books and junior readers.
For one thing, this makes it impossible to pronounce place names on road signs correctly unless you already know the name or can read the characters.
That's a great example of a bad sentence. Did they receive five weeks of funding, or less than 47 weeks?
Uh-oh, have to read TFA to find out...
Raytracing scales O(pixels) and O(ln(complexity)). Rasterization is relatively constant in the number of pixels, and O(complexity).
That being the case, and contrary to what many are claiming, ray tracing looks to be the loser in terms of computational complexity...
If we assume scene complexity increases at most in proportion to pixel count (seems reasonable, as we don't want to waste effort describing sub-pixel details), we get:
Ray tracing: O(n log n)
Rasterization: O(n)
Hey, you managed to use both "your" and "you're" both correctly and incorrectly in one post!
I think you must be taking the idea of stochastic grammar somewhere it doesn't belong...
TFA has two points to make:
1. Bad science has been used in supporting claims of Dvorak's superiority
2. There is evidence that *re*-training in Dvorak isn't cost effective
There's nothing there to suggest that Dvorak isn't superior to QWERTY, at least for those who type Dvorak from the get-go.
Hands up how many Dvorak advocates actually advocate learning QWERTY first, and only then switching to Dvorak? Just for kicks? I thought not.
No 'myth' is being 'debunked' here.
Agreed, being able to type a bit faster is nice, but the main benefit of Dvorak for heavy usage is the improved ergonomics.
Re-training in Dvorak might not be cost-effective for the employer, but how about the poor employee (and their health provider) who gets their wrists f*cked from typing QWERTY all day?
I'm not going to advocate Dvorak for new typists though, because much like QWERTY it was designed for manual typewriters, and computer keyboards are different.
Dvorak tends to split digraphs across the two hands. This is good for mechanical typewriters that require force from the wrist. For computer keyboards, it's faster to type pairs of adjacent letters on the same hand.
Maybe something like 'Snaither' here...
1 2 3 4 5 6 7 8 9 0 [ ]
L F O P J Z M D Y W / =
S N A I . , T H E R -
G U Q B ' ; V K X C
> Other people hate it because there's nobody else to kick the bucket with.
If you're all feeling like topping yourselves, best to take the day off!
What I wanna do is set up a honey pot dresser drawer that's wired to a silent alarm, maybe one that sends me an SMS and activates surveillance cameras.
The contents of the honey pot drawer would of course include something of value with a homing beacon concealed in it.
(But maybe I won't bother now everyone knows how secure my house is.)