This reminds me of a story about leaded plastic. An American company had plastic rules produced in China, and the tested positive for lead but not to any immediately dangerous degree unless you were sticking them in your mouth. The American company was furious at the Chinese manufacturer and questioned why they would make such a thing. The Chinese manufacturer, confused, asked "why would any child stupid enough to put a ruler in their mouth be using a ruler in the first place?".
These magnets are basically the same thing - they are dangerous and they shouldn't be near little kids. You hide your knives, lighters, batteries, and medicine from them so why should this be any different. The article notes a case where it was in a construction kit - even the toys my 3 year old plays with I wouldn't let near a baby so it's their own fault for not managing their environment. I mean come on, the kids swallowed not one but two of them on -different- days.
I see your point but you need to realize some OSS projects can be used to provide services which you can make money off of. I participate in one such project and I use it in a service I provide to clients. That could include embeded operating systems as well, just look at iTRON, TOPPERS, Arduino, etc.
But... in the case of WebOS I'd say you are absolutely correct. At this point the best HP could really hope for is the code based be used as a reference to implement attractive features in some other system. May as well call the new OSS version OrphanOS.
Well, a very large earthquake, being steeped in contaminated water for almost a year, hydrogen explosions, large metal vessels being knocked over... there's a lot of ways those cracks could have formed during and after the incident.
Or at least the suspicion there may be seepage through cracks in the foundation. It was in the news quite a while ago, I guess they just released some numbers now and that's what the article was referring to? It's not the first "spill" either, one of the pools overflowed and some water was released into the ocean.
Because bug testers cost money and they don't always find all the bugs. I've never ran one of these larger reporting systems or even been part of the administration process of one but many big projects have them. EG: FireFox, Ubuntu, etc.
There are reporting systems that let you manage priorities and to triage them/assign developers to fix them etc. On small projects it's overkill I think though.
That sounds like a nice scheme, I'd love to see your implementation or if you're using a library or something I'd like to know the name. Is there an option for user comments?
Precisely, or better yet include let them add a comment to the dump and ask them to recall what they did just before the dump occurred. If you can match up their recalled actions to the dump information I've found it makes things a lot easier.
On a related note anybody else remember "core" files stuffing up your home directory?
Yes I'm absolutely sure it was in UTF-8. It was for inline UTF-8 in C++ and any "non-UTF-8" characters after a UTF-8 character would be garbled or if it was the last character in a string that last character itself would be garbled or there would be trailing ASCII. In addition comments would screw things up but that was easily fixed by adding a few spaces after each comment line.
As for the BOM it was being added in automatically, so as you point out it was removed from saved.h/.cpp files by default that is something that happened after the project I was talking about. And now that you mention it I believe VS2010 had just come out as they were still finishing moving the code from an older version when I was called into the project. To this day I actually have no idea if there were issues with the UTF-8 in older VS versions so maybe it was just a post-release bug. Perhaps that's an indicator they did fix the problem. I'll actually be porting an application to Windows some time early next year and this is an issue I was worried about so I hope they did.
Interesting about PowerShell, I have to admit that looks like fairly nice syntax. No SSH is the biggest hangup for me because I use SSH constantly and not just for remote shell sessions, and as you may have gathered I don't work on English systems so much so Cygwin SSH barely cuts it and Putty doesn't do everything I need it to. I've honestly never found a single solution on Windows that worked for me.
I haven't used VS for about a year now but the last time I did it was UTF-8 that we were having issues with. I was working on some localization for a client, in the end we had to actually export strings to an external file because the UTF-8 codepage in VS was completely incompatible with Linux and OSX. We couldn't figure out a method to make it work and we called MS. It was their solution to move the strings out to external files. Before you ask, no this was not a BOM issue, though including a BOM would have made the files unusable on other systems in and of itself. Unless things have changed in the last year (I hope they have) I would not count VS as "supporting" UTF-8.
I agree with the [g]vim part, but I've never found PowerShell to be particularly useful and (correct me if I'm wrong) standard tools like ssh, grep, etc. aren't included. I also often deal with UTF-8 and absolutely everything in Windows (7 included) regarding UTF-8 is a disaster - especially when it comes to the terminal - and cygwin does a terrible job with it as well.
While I personally dislike working on OSX Hatta does have a point in that you have a functional terminal and all the tools you should have (ssh, decent shells, scripting facilities and editors) on a basic system. The fact is in Windows you need to install Cygwin for that and even then it's a mess. Of course for people writing documents and editing photos that's all completely unrelated, but for people like me who hate the standard interface it's nice to know I can just open a terminal and have the facilities to do what I need to and be done with it.... I'd also spend 10 minutes trying to get rid of that damn dock because it bugs the hell out of me and that's the same 10 minutes I'd spend installing Cygwin...
Actually I very much agree with him, and I'm a user. I feel like OSX is trying to make me do things with mittens on, the freedom of my bare hands obstructed with a warm fuzzy enveloping layer. And I absolutely disagree that the Ribbon is a better interface as well, I want to know exactly where things are and have them there all the time even if they aren't related to the current context I'm working in.
I am not however rejecting "new" interfaces - now that there's an extension to add a taskbar to GNOME 3 (shell) and after tweaking it a bit I feel like I can use it more efficiently than GNOME 2 now, and like it. I'm an old user, and though I resisted a bit I'm all for change and I'm enjoying change that lets me take more control and work more efficiently.
As for the Ribbon and new users, I have clients who hated it so much that when I showed them OO/LibreOffice they immediately switched. That says a lot if you ask me.
Yeah, you wrote a bunch of comments and glancing over them I didn't see anything that immediately points to why parent is incorrect. Let me teach you some probability here: 1. Probability people will assume you are an opinionated jerk with an agenda with a post like that: HIGH 2. Probability the average Slashdot reader would actually bother to look up, read, and consider how each of your posts in response to this article correlates the parents post and in what ways it proves he is incorrect: LOW
What's more sad is I have minor Discalculia (and Dyslexia) and I blew through the sample questions without a calculator in about 2 minutes. If you need a calculator for even one of those questions I think you have a problem.
I send PDFs usually, if I'm sending a document in an editable format it's because whoever I'm working with (a client for example) is going to add something and send it back to me - in which case there isn't going to be a lot of formatting on the document to begin with. Linux breaking in an upgrade is an odd thing to say too, I've had some quirks but never anything serious and certainly nothing show-stopper level. As for users are you talking about clients? Most of my current clients have mixed systems and it's been that way for a long time, so I write every desktop application in cross-platform frameworks/languages anyway.
Now if you are a MS shop writing C# and.net I could understand your point, but we are most certainly not.
THIS! We were mainly OSS to begin with and that saved us, and we actually had all the boxes and receipts for everything else. But in the end after the BSA dropped their case we wiped everything and now every corporate machine is 100% OSS. I honestly hope they come after us again so we can just laugh at them. Fuckers.
In Japan as soon as you contact the service provider they remotely lock the phone, start tracking it, and if you've reported it stolen they report its position to the police.
Hah, I've had similar experiences with Perl, and from Perl Python would be a dream for sure. As for older Ruby (when it was first released) it wasn't that great, and the upcoming Ruby 2.0 is going to change a lot of things too. I personally like that Ruby is a constantly evolving language, but I could also see how that in and of itself could put people off, especially anyone who prefers or needs a high level of consistency and stability in a language.
Oh yes, I remember pouet but I've never posted. I've never actually participated in a full demo either, being in the industry at the time I was bound by some scary contracts and I didn't want to risk anything.
As for GCC and binary size there are a few tricks, like actually defining your own C runtime (easier than you'd think) etc. but in the end it's always a little tweaky so binary sizes are hard to predict. And "OO" C with structs and function pointers was how a lot of people handled things, we even had different names for different styles where I worked and we did everything from entities to states to menu systems with them. If you're doing flow control with things try having a loop that just calls the same function pointer over and over again and have state functions set that pointer on the tail end - it can save you switch statements and long conditional logic.
I like how you even used sed to switch up the slashes and break down long names. Now to sneak this into someones bashrc and catch their face when they login...
I get your points and I thought I'd acknowledge I understood your standpoint. But I honestly think you completely missed or just chose to ignore my points and you don't seem like the person who would care to so just let's just stop here. I really don't care to bicker on the internet.
The Scene.org thing is awesome though. The low level stuff I was talking about was mainly on a particular piece of popular gaming hardware which I wrote engines on. I would frequent scene.org for tips and tricks and I can honestly say it was because of that I was able to pull off some amazing stuff (including a very functional 3D engine and pull off quite a few awesome effects) on an extremely limited device.
Consistent object model: list.length vs. len(list)
Consistent with what? if we're going to be pedantic here, it's list.__len__(). The len() function calls the internal property where available (ie. you specifically wrote one). A special data type doesn't need to actually implement __len__ to do something sane with len. Or rather, it's to make special data types (eg. fast C arrays) consistent with normal ones. It's __len__ rather than 'length' because you may want to define the real word argument as something else without losing consistency. Auto methods which have a built in language are all named like that. For consistency.
I'm not sure how you missed the point there. Being able to extend objects almost universally is something I think most people would consider a very nice feature. How you can argue that having some universal external function (which you'd have to add overrides for for new types) is better? I guess if you hate having class methods perform functions on their own objects you'd have a point... but then I'd assume you hate languages like C++ and Java as well?
Enumeration is just another method of objects that are enumerable. list.each {|item|.... }
Or list.each_item, or list.try_convert, or all of the other built in lambda functions, all of which can be replaced with a list comprehension, which can do a lot more without learning more than one syntactic concept.
simple.
ugly, and leaves you reading the reference guide for the array class every time you're trying to find the method you're after because there's way too many.
Then use the ones you know, and when they don't work out for you or you think you can do it cleaner with a different one look it up.
Method calls that return a boolean should have a question mark
Why? Do methods that return an into have a big letter I? Do string methods have an S? What about methods that return a metaclass? That's not consistent. It's syntactic litter.
Why? Because it looks nice and more like natural language. Just like using Hungarian Notation it's a syntax flag that lets people know what's going on without actually looking at the internals of the function. Or would you perfer something like "B_IsThisValid(thing)" than "thing.valid?" ?
Everything returns an object including if/else statements.
If/else is flow control, not a function. Only objects which can meaningfully be applied to an lvalue context return an object in python. Not sure what crazy indentation you would use to embed an if/else statement into somewhere you would use that whilst trying to keep it readable. That's simply not useful. By adding the whopping extra word "return" within a block, you can make it clear that you're returning a value, and more importantly, which value, rather than the implied one which just happens to be the result of the last statement
You can use return in Ruby, and "everything returns an object" is just part of everything in Ruby basically being an object. Actually reading your comment there I'm increasingly unsure you've actually coded in Ruby, or if you actually have I have a feeling you just didn't grasp some of the core paradigms.
and open classes make it easy to do
Open classes (or "duck typing", or "monkey patching" as it's also affectionately known) make it easy to modify a perfectly well documented class by changing its functionality, then wondering why upgrading a library broke it. You can actually do this in python just as easily, but the difference is that while ruby programmers think it's the best thing EVAR, hop over to an equivalent python list and you can find plenty of essays abou
Slashdot Mirror
This reminds me of a story about leaded plastic. An American company had plastic rules produced in China, and the tested positive for lead but not to any immediately dangerous degree unless you were sticking them in your mouth. The American company was furious at the Chinese manufacturer and questioned why they would make such a thing. The Chinese manufacturer, confused, asked "why would any child stupid enough to put a ruler in their mouth be using a ruler in the first place?".
These magnets are basically the same thing - they are dangerous and they shouldn't be near little kids. You hide your knives, lighters, batteries, and medicine from them so why should this be any different. The article notes a case where it was in a construction kit - even the toys my 3 year old plays with I wouldn't let near a baby so it's their own fault for not managing their environment. I mean come on, the kids swallowed not one but two of them on -different- days.
I see your point but you need to realize some OSS projects can be used to provide services which you can make money off of. I participate in one such project and I use it in a service I provide to clients. That could include embeded operating systems as well, just look at iTRON, TOPPERS, Arduino, etc.
But... in the case of WebOS I'd say you are absolutely correct. At this point the best HP could really hope for is the code based be used as a reference to implement attractive features in some other system. May as well call the new OSS version OrphanOS.
Well, a very large earthquake, being steeped in contaminated water for almost a year, hydrogen explosions, large metal vessels being knocked over... there's a lot of ways those cracks could have formed during and after the incident.
Or at least the suspicion there may be seepage through cracks in the foundation. It was in the news quite a while ago, I guess they just released some numbers now and that's what the article was referring to? It's not the first "spill" either, one of the pools overflowed and some water was released into the ocean.
Excellent, I'll take a look when I can. Thanks.
Anything multi-platform? That looks like it's Windows only.
Because bug testers cost money and they don't always find all the bugs. I've never ran one of these larger reporting systems or even been part of the administration process of one but many big projects have them. EG: FireFox, Ubuntu, etc.
There are reporting systems that let you manage priorities and to triage them/assign developers to fix them etc. On small projects it's overkill I think though.
That sounds like a nice scheme, I'd love to see your implementation or if you're using a library or something I'd like to know the name. Is there an option for user comments?
Precisely, or better yet include let them add a comment to the dump and ask them to recall what they did just before the dump occurred. If you can match up their recalled actions to the dump information I've found it makes things a lot easier.
On a related note anybody else remember "core" files stuffing up your home directory?
Yes I'm absolutely sure it was in UTF-8. It was for inline UTF-8 in C++ and any "non-UTF-8" characters after a UTF-8 character would be garbled or if it was the last character in a string that last character itself would be garbled or there would be trailing ASCII. In addition comments would screw things up but that was easily fixed by adding a few spaces after each comment line.
As for the BOM it was being added in automatically, so as you point out it was removed from saved .h/.cpp files by default that is something that happened after the project I was talking about. And now that you mention it I believe VS2010 had just come out as they were still finishing moving the code from an older version when I was called into the project. To this day I actually have no idea if there were issues with the UTF-8 in older VS versions so maybe it was just a post-release bug. Perhaps that's an indicator they did fix the problem. I'll actually be porting an application to Windows some time early next year and this is an issue I was worried about so I hope they did.
Interesting about PowerShell, I have to admit that looks like fairly nice syntax. No SSH is the biggest hangup for me because I use SSH constantly and not just for remote shell sessions, and as you may have gathered I don't work on English systems so much so Cygwin SSH barely cuts it and Putty doesn't do everything I need it to. I've honestly never found a single solution on Windows that worked for me.
I haven't used VS for about a year now but the last time I did it was UTF-8 that we were having issues with. I was working on some localization for a client, in the end we had to actually export strings to an external file because the UTF-8 codepage in VS was completely incompatible with Linux and OSX. We couldn't figure out a method to make it work and we called MS. It was their solution to move the strings out to external files. Before you ask, no this was not a BOM issue, though including a BOM would have made the files unusable on other systems in and of itself. Unless things have changed in the last year (I hope they have) I would not count VS as "supporting" UTF-8.
I agree with the [g]vim part, but I've never found PowerShell to be particularly useful and (correct me if I'm wrong) standard tools like ssh, grep, etc. aren't included. I also often deal with UTF-8 and absolutely everything in Windows (7 included) regarding UTF-8 is a disaster - especially when it comes to the terminal - and cygwin does a terrible job with it as well.
While I personally dislike working on OSX Hatta does have a point in that you have a functional terminal and all the tools you should have (ssh, decent shells, scripting facilities and editors) on a basic system. The fact is in Windows you need to install Cygwin for that and even then it's a mess. Of course for people writing documents and editing photos that's all completely unrelated, but for people like me who hate the standard interface it's nice to know I can just open a terminal and have the facilities to do what I need to and be done with it. ... I'd also spend 10 minutes trying to get rid of that damn dock because it bugs the hell out of me and that's the same 10 minutes I'd spend installing Cygwin...
Actually I very much agree with him, and I'm a user. I feel like OSX is trying to make me do things with mittens on, the freedom of my bare hands obstructed with a warm fuzzy enveloping layer. And I absolutely disagree that the Ribbon is a better interface as well, I want to know exactly where things are and have them there all the time even if they aren't related to the current context I'm working in.
I am not however rejecting "new" interfaces - now that there's an extension to add a taskbar to GNOME 3 (shell) and after tweaking it a bit I feel like I can use it more efficiently than GNOME 2 now, and like it. I'm an old user, and though I resisted a bit I'm all for change and I'm enjoying change that lets me take more control and work more efficiently.
As for the Ribbon and new users, I have clients who hated it so much that when I showed them OO/LibreOffice they immediately switched. That says a lot if you ask me.
Yeah, you wrote a bunch of comments and glancing over them I didn't see anything that immediately points to why parent is incorrect. Let me teach you some probability here:
1. Probability people will assume you are an opinionated jerk with an agenda with a post like that: HIGH
2. Probability the average Slashdot reader would actually bother to look up, read, and consider how each of your posts in response to this article correlates the parents post and in what ways it proves he is incorrect: LOW
What's more sad is I have minor Discalculia (and Dyslexia) and I blew through the sample questions without a calculator in about 2 minutes. If you need a calculator for even one of those questions I think you have a problem.
niftydude is currently rated 4, Funny - but the sad thing is he's technically correct.... this is Insightful, or at lest Informative.
I send PDFs usually, if I'm sending a document in an editable format it's because whoever I'm working with (a client for example) is going to add something and send it back to me - in which case there isn't going to be a lot of formatting on the document to begin with. Linux breaking in an upgrade is an odd thing to say too, I've had some quirks but never anything serious and certainly nothing show-stopper level. As for users are you talking about clients? Most of my current clients have mixed systems and it's been that way for a long time, so I write every desktop application in cross-platform frameworks/languages anyway.
Now if you are a MS shop writing C# and .net I could understand your point, but we are most certainly not.
THIS! We were mainly OSS to begin with and that saved us, and we actually had all the boxes and receipts for everything else. But in the end after the BSA dropped their case we wiped everything and now every corporate machine is 100% OSS. I honestly hope they come after us again so we can just laugh at them. Fuckers.
In Japan as soon as you contact the service provider they remotely lock the phone, start tracking it, and if you've reported it stolen they report its position to the police.
Hah, I've had similar experiences with Perl, and from Perl Python would be a dream for sure. As for older Ruby (when it was first released) it wasn't that great, and the upcoming Ruby 2.0 is going to change a lot of things too. I personally like that Ruby is a constantly evolving language, but I could also see how that in and of itself could put people off, especially anyone who prefers or needs a high level of consistency and stability in a language.
Oh yes, I remember pouet but I've never posted. I've never actually participated in a full demo either, being in the industry at the time I was bound by some scary contracts and I didn't want to risk anything.
As for GCC and binary size there are a few tricks, like actually defining your own C runtime (easier than you'd think) etc. but in the end it's always a little tweaky so binary sizes are hard to predict. And "OO" C with structs and function pointers was how a lot of people handled things, we even had different names for different styles where I worked and we did everything from entities to states to menu systems with them. If you're doing flow control with things try having a loop that just calls the same function pointer over and over again and have state functions set that pointer on the tail end - it can save you switch statements and long conditional logic.
I like how you even used sed to switch up the slashes and break down long names. Now to sneak this into someones bashrc and catch their face when they login...
I get your points and I thought I'd acknowledge I understood your standpoint. But I honestly think you completely missed or just chose to ignore my points and you don't seem like the person who would care to so just let's just stop here. I really don't care to bicker on the internet.
The Scene.org thing is awesome though. The low level stuff I was talking about was mainly on a particular piece of popular gaming hardware which I wrote engines on. I would frequent scene.org for tips and tricks and I can honestly say it was because of that I was able to pull off some amazing stuff (including a very functional 3D engine and pull off quite a few awesome effects) on an extremely limited device.
Consistent with what? if we're going to be pedantic here, it's list.__len__(). The len() function calls the internal property where available (ie. you specifically wrote one). A special data type doesn't need to actually implement __len__ to do something sane with len. Or rather, it's to make special data types (eg. fast C arrays) consistent with normal ones. It's __len__ rather than 'length' because you may want to define the real word argument as something else without losing consistency. Auto methods which have a built in language are all named like that. For consistency.
I'm not sure how you missed the point there. Being able to extend objects almost universally is something I think most people would consider a very nice feature. How you can argue that having some universal external function (which you'd have to add overrides for for new types) is better? I guess if you hate having class methods perform functions on their own objects you'd have a point... but then I'd assume you hate languages like C++ and Java as well?
Or list.each_item, or list.try_convert, or all of the other built in lambda functions, all of which can be replaced with a list comprehension, which can do a lot more without learning more than one syntactic concept.
ugly, and leaves you reading the reference guide for the array class every time you're trying to find the method you're after because there's way too many.
Then use the ones you know, and when they don't work out for you or you think you can do it cleaner with a different one look it up.
Why? Do methods that return an into have a big letter I? Do string methods have an S? What about methods that return a metaclass? That's not consistent. It's syntactic litter.
Why? Because it looks nice and more like natural language. Just like using Hungarian Notation it's a syntax flag that lets people know what's going on without actually looking at the internals of the function. Or would you perfer something like "B_IsThisValid(thing)" than "thing.valid?" ?
If/else is flow control, not a function. Only objects which can meaningfully be applied to an lvalue context return an object in python. Not sure what crazy indentation you would use to embed an if/else statement into somewhere you would use that whilst trying to keep it readable. That's simply not useful. By adding the whopping extra word "return" within a block, you can make it clear that you're returning a value, and more importantly, which value, rather than the implied one which just happens to be the result of the last statement
You can use return in Ruby, and "everything returns an object" is just part of everything in Ruby basically being an object. Actually reading your comment there I'm increasingly unsure you've actually coded in Ruby, or if you actually have I have a feeling you just didn't grasp some of the core paradigms.
Open classes (or "duck typing", or "monkey patching" as it's also affectionately known) make it easy to modify a perfectly well documented class by changing its functionality, then wondering why upgrading a library broke it. You can actually do this in python just as easily, but the difference is that while ruby programmers think it's the best thing EVAR, hop over to an equivalent python list and you can find plenty of essays abou