Webforms is the anti-christ to any competent web UI developer. All thin-client solutions are. Since the DOM API became reasonably well-supported around 2000, reloading the page to swap static lists for a paired combo box widget was never a competent or necessary thing to do. Loading dynamic data in via ajax moves pretty damn fast nowadays if you keep processing of data to a minimum on the back-end.
If nobody would use JavaScript given alternatives, then why is it spreading to desktop-os, server-side, and cross-platform mobile apps (to be fair - it's not the only choice I'd make for mobile at this time - canvas support blows and web views are still ragged bleeding edge)? There were other options in the '90s, although IIRC I think you had to download interpreters. And of course Java Applets, VB, Flash, etc... At the end of the day the browser vendors know what every UI dev knows. A primarily LISP/Scheme-inspired language like JS is an excellent design for the task. It's easy for rookies to pick up and start doing basic stuff with it but there's also a powerful language worthy of study there that happens to be a good fit for the roles its had so far if you take it seriously.
What people fail to understand, is that your typical experienced UI dev tends to have had a lot more exposure to other languages than your typical web server-side dev, most of whom can't even be bothered to become functionally literate in HTML beyond basic syntax which they new from XML anyway and often have no idea how CSS works. Those of us moving on to generalist territory aren't just picking Node.js for a lack of basis for comparison or because we're afraid to learn anything other than JS. I was actually a little wary of Node at first until I tried it. I've always enjoyed Python when I needed it for file-system type stuff and I loved Django (but sadly could never find work with anybody who was using it), Ruby wouldn't suck if it had a decent interpreter but I've never understood the excitement over Rails which I find 'meh, C# can be tolerable with MVC but why you would pay for it with so many superior alternatives available is beyon me, every experience I've ever had with Java-only teams or the disaster code they leave behind has been completely miserable (the language is a little crude but usable), and I'm a little too turned off by PHP's lack of consistency in API to take it especially seriously for my own use but I respect a good PHP dev.
I probably like Node as much as I do for reasons typical to JS devs.
* Takes 5 minutes to get it up and running on any platform - no surprise there. We're used to executing code right out of a console or by hitting "refresh" in a browser. Having to fiddle with XML config before you can even compile, servers with really pointless default settings, or a cluster-fucked code base that gets into a fight between Ant, Eclipse AND Maven because some tools wanted every technology they could think of on their resume is not something I would expect to run into with Node.js. You install it on Mac, Windows, Linux, it runs. Anything expecting a great deal more effort than that just to execute code tends to get tossed.
* Async event-driven architecture in a familiar format. All that concurrency stuff everybody gets so excited about is just second nature to anybody who's been handling event driven UI code for years. As are a lot of aspects of functional programming. We're not almost there. We don't need to attend workshops. It's just what we've been doing for years with a language that serves those needs well.
* Whoever's benchmarks you're reading, I think it's safe to say Node's performance is strong at this point. Barring stuff from people who still don't quite get the concept of needing to keep things async with the JavaScript.
* C/C++ bindings - two things JS doesn't have, design for more advanced math handling and performance that absolutely hands-down kicks absolutely everybody's ass, easily glued to. It's a nice tag-team. All the architecture/implementation advantages of an extremely flexible scripting language with C right there to burn through any computationally intensive tasks for a performance advantage or to handle scenarios where static types can be advantageous. JS here and C there will always appeal to me way more than Java or.NET and all the baggage they bring.
* It's straightforward - It's mostly just a C event loop with a not-gargantuan fil
Oh, FFS, stop crying and learn how to take responsibility for flow of data. If Java and C# devs at the median level knew the first thing about OOP vs. procedural they wouldn't piss their pants every time they saw a dynamically typed scripting language. I moved from client-side work to becoming more of a generalist and I am completely blown away by how moronic the vast majority of server-side code I've been exposed to is. Absurdly deep inheritance schemes. Getters/setters everywhere. Design patterns that make no sense in the context they've been placed in. How the hell is OOP any different from crap-procedural code when you have 25 methods from 25 different classes and objects all mutating the same data? It's just function spaghetti with redundant class syntax.
Type systems are about design tradeoffs (performance and compiler debug convenience vs. flexibity), not whether the language is closer to the one-true-way to do it right. Anybody who believes static types makes their code more robust is lying to themselves. If static typing protects you from anything it's from learning how not to write illegible, unmaintainable shit code that can only be tended to by using IDE crutches to work stupid faster.
In C/C++ I wouldn't have it any other way than static. When I'm writing UI and higher-level architecture, I'll take my variable length args, first-class functions, closures, extreme object mutability (but still not lacking for encapsulation), because when you actually have a little bit of discipline and attention to detail, you can get a lot more done with a lot less code and that is always easier to maintain, reuse and modify in the long haul.
It doesn't matter what the language was. Lack of proper support for the DOM standard on IE's part was the problem, not JavaScript. The ECMA spec has actually been very evenly supported since IE5. Always behind on the latest of course but at least we didn't have to polyfill native language methods in to replace stupid proprietary MS shenanigans for the core language.
But the last version of Netscape Navigator was closer to full compliance with CSS2 and the DOM specs than IE8 was. It took those assholes 10 !@#$ing years to add table-display properties. We've been normalizing the entire event system for IE's sake for a decade too until IE9 came out.
All of this by the way is why JS ended up reigning supreme on the client-side web. Nothing in popular use normalizes or handles UI concerns as well as JS does.
Okay, that's pretty yucky, but also basically DWI speed traps. This is cops pulling people over for no reason so that a private contracting firm can make 7.9 million dollars conducting a survey that will be statistically suspect at best. Cops doing something other than their job to interrupt people's lives not even for a DWI fishing expedition but acting as corporate minions.
All non-free software is a rip-off and a scam, and I am convinced we should be able to prove it.
I agree with the majority of your post but I've always had an issue when people take it to this extreme. Is it maybe reasonable for a game developer to not give anybody and everybody permission to distribute their single-player PC-only game whose design assets alone cost millions to create? Is it even responsible to distribute the source of a web app that handles confidential user information? Is anybody who works for a non-free software company collaborating in a scam?
The Free Software Foundation kicks ass but if all the software of the world was free (yes I mean not as in beer but there is no real distinction for certain types of apps) there'd be little motivation for a lot specific types of software to have been made at the level of quality we've enjoyed in the first place.
Software isn't always just code. It's often composed of intellectual properties and other concerns that can't be protected if distributed freely. Actors like getting royalties for having their likenesses spread all over creation. Artists don't like having their distinctive works appropriated without being credited.
And what is a web-site but an app that's being distributed from one source only?
Coming from a web background, the idea of building a non-Free app using Free tools doesn't bother me one wit as long as you give back by contributing useful tools you developed in the process or giving feedback/support on the ones you used.
IMO, free and non-free apps are just a part of the same ecosystem. An Office Suite that let you rapidly port data for use in its various apps would have been a loonnnng time coming originating as an open-source endeavor. But there was serious money to be made there, which is a great motivator to find a way to get multiple teams on very different projects to learn to play nice together. This is not something open source communities collaborating between projects are necessarily famous for.
But as an alternative, a free version of an office suite also happened a lot sooner because MS has been a total douchebag about its software by turning interoperability into handcuffs. As long as OO has any kind of marketshare MS can only take the douchebaggery so far and I'd argue that free software has already directly and indirectly pushed them to make some positive changes. Everybody wins. As free tools and communities mature, non-free vendors will be forced to respond in kind and I suspect we'll all continue to win even more.
So don't see it as an adversarial thing. It's a relationship. A really fucked up dysfunctional one but that happens to produce a decent kid mostly by accident.
Just because open source devs seems smarter, doesn't meant they're paid any better. Hell, I'll bet they're typically paid less. A consequence of actually enjoying the work and not having to work as late perhaps.
Any office having trouble with a docx file in 2013 without the use of MSO is backward. Haven't had a problem opening in Google docs in years. Pretty sure OO's had that covered for a while too.
Also, winmail.dat is 100% a Microsoft-manufactured problem. They translate certain types of attachments to a proprietary Outlook format for security concerns real or imagined on SEND rather than simply buffering with said Outlook shenanigans on receipt but not resend so nobody else has to put up with Outlook's slow-ass dinosaur app BS but Outlook users. To be fair to MS, I suspec the Microsoft TwoFace coin landed on the incompetent moron side rather than the one featuring the slightly less incompetent devious asshole that day or they'd be doing the winmail.dat thing with a lot more file formats.
But fortunately that problem's already been compensated for with this plugin for Thunderbird. Found it at the top of my first Google. It's been out for a year. Maybe you should let your IT guys know before they get locked into... oh... too late aren't I?
Well at least when your users start pissing off some Thunderbird-using company that rightly identifies emails from clueless Outlook users as the problem, they can pass on that link so they don't have to pay money for the privilege of becoming a part of the same problem too.
As a web UI developer I can promise you there have DEFINITELY been no such shenanigans with Internet Explorer which has blown goats honestly at least since 2000 and probably since IE first actually had competition.
I gathered from the article that the whole point was more, "Don't try to justify it in terms of savings or you will fail the second you hit a snag and it seems cheaper/easier to just go back than retrain anybody or do some research into some alternative you need."
Having grown up in a time when personal computers were just starting to get popular, you start to realize that some people will never be more than Foundation trilogy-style technopriests due to a deep and abiding fear of things they don't understand that literally paralyzes them into monkey see, monkey do behavior where technology is concerned. They are taught to click that icon, then that file thing in the gray bar and then "open" and then that folder name, then that file. Change that folder name or move it into subcategory somewhere right in front of their eyes and holy cow there's a problem. Move the entire file menu into a pretty windows icon button and it's brown trousers time. I'm not sure if it never occurs to them that there's some kind of a generalized process or scheme that they're following that could actually help them if they understood it at just one level of abstraction deeper or if they're just not actually capable of understanding things that way.
I see less of it now but still run into it even in younger people who've been around tech all their lives.
It's kind of like people who never truly learn to read properly. Beyond basic consonant use, words to them are mostly a series of letters they associate with words they speak rather than there being any real pattern to how the sequences of letters actually inform the sounds most of the time. My last name for instance is similar in vowel/consonant construction to "hatter." I'd say about half the time when somebody has to speak it out loud they screw up their eyes and think and then what comes out is "hater." It makes my wife, who is a human Oxford-English dictionary (including pronunciation keys) batshit.
The only way this surveillance crap will ever end well is if everybody gets to look at everybody's data. If I can check on where my political leaders, government officials, cops, and corporate leadership is at any time and see what they've been doing online, I have no objection to them doing the same to me or anybody else. Until that day, it's a problem. Because then it's not about whether I have anything to hide but rather whether they'd like to be able to misbehave without any real risk of me even trying to do anything or say anything about it.
At one point the website would require you to navigate 8+ slowly-loading, cookie-dropping pages after you'd pulled the trigger on wanting to buy something. I once witnessed the legal department getting its 2 cents in and forcing change on some logos that it didn't like the look of. Purely aesthetic concern. It is the single-most dysfunctional corporate culture I've ever encountered.
At the time, they were running the show with 100s of offshore devs whose programming literacy I would place on a scale of non to bordering on competent. No version control. I am dead !@#$ing serious.
My vote's on corporations being allowed to buy votes with impunity, when what they and their lobbyists minions really deserve is to be drawn and quartered. I may not be entirely objective on the issue but I think that may have something to do with years of frustration of watching joe public getting beaten with a carrot and told to eat the stick and liking it.
Uh... the NSA HAS been so inclined. Lots of protesters getting accused of terrorism in recent years. We've got three kids in Chicago that have been in jail since the NATO summit. It took several months for the actual charges to be brought against them. They can do that now. Throw you in jail with no charges and come up with 'em later. If they follow the usual pattern, they'll drop the charges in a year or two when absolutely everybody has stopped paying attention. I'm not some Occupy kid. The police reaction to the protests was downright spooky. That the NSA was involved in a domestic affair is deeply disturbing.
Slashdot Mirror
Webforms is the anti-christ to any competent web UI developer. All thin-client solutions are. Since the DOM API became reasonably well-supported around 2000, reloading the page to swap static lists for a paired combo box widget was never a competent or necessary thing to do. Loading dynamic data in via ajax moves pretty damn fast nowadays if you keep processing of data to a minimum on the back-end.
If nobody would use JavaScript given alternatives, then why is it spreading to desktop-os, server-side, and cross-platform mobile apps (to be fair - it's not the only choice I'd make for mobile at this time - canvas support blows and web views are still ragged bleeding edge)? There were other options in the '90s, although IIRC I think you had to download interpreters. And of course Java Applets, VB, Flash, etc... At the end of the day the browser vendors know what every UI dev knows. A primarily LISP/Scheme-inspired language like JS is an excellent design for the task. It's easy for rookies to pick up and start doing basic stuff with it but there's also a powerful language worthy of study there that happens to be a good fit for the roles its had so far if you take it seriously.
What people fail to understand, is that your typical experienced UI dev tends to have had a lot more exposure to other languages than your typical web server-side dev, most of whom can't even be bothered to become functionally literate in HTML beyond basic syntax which they new from XML anyway and often have no idea how CSS works. Those of us moving on to generalist territory aren't just picking Node.js for a lack of basis for comparison or because we're afraid to learn anything other than JS. I was actually a little wary of Node at first until I tried it. I've always enjoyed Python when I needed it for file-system type stuff and I loved Django (but sadly could never find work with anybody who was using it), Ruby wouldn't suck if it had a decent interpreter but I've never understood the excitement over Rails which I find 'meh, C# can be tolerable with MVC but why you would pay for it with so many superior alternatives available is beyon me, every experience I've ever had with Java-only teams or the disaster code they leave behind has been completely miserable (the language is a little crude but usable), and I'm a little too turned off by PHP's lack of consistency in API to take it especially seriously for my own use but I respect a good PHP dev.
I probably like Node as much as I do for reasons typical to JS devs.
* Takes 5 minutes to get it up and running on any platform - no surprise there. We're used to executing code right out of a console or by hitting "refresh" in a browser. Having to fiddle with XML config before you can even compile, servers with really pointless default settings, or a cluster-fucked code base that gets into a fight between Ant, Eclipse AND Maven because some tools wanted every technology they could think of on their resume is not something I would expect to run into with Node.js. You install it on Mac, Windows, Linux, it runs. Anything expecting a great deal more effort than that just to execute code tends to get tossed.
* Async event-driven architecture in a familiar format. All that concurrency stuff everybody gets so excited about is just second nature to anybody who's been handling event driven UI code for years. As are a lot of aspects of functional programming. We're not almost there. We don't need to attend workshops. It's just what we've been doing for years with a language that serves those needs well.
* Whoever's benchmarks you're reading, I think it's safe to say Node's performance is strong at this point. Barring stuff from people who still don't quite get the concept of needing to keep things async with the JavaScript.
* C/C++ bindings - two things JS doesn't have, design for more advanced math handling and performance that absolutely hands-down kicks absolutely everybody's ass, easily glued to. It's a nice tag-team. All the architecture/implementation advantages of an extremely flexible scripting language with C right there to burn through any computationally intensive tasks for a performance advantage or to handle scenarios where static types can be advantageous. JS here and C there will always appeal to me way more than Java or .NET and all the baggage they bring.
* It's straightforward - It's mostly just a C event loop with a not-gargantuan fil
Oh, FFS, stop crying and learn how to take responsibility for flow of data. If Java and C# devs at the median level knew the first thing about OOP vs. procedural they wouldn't piss their pants every time they saw a dynamically typed scripting language. I moved from client-side work to becoming more of a generalist and I am completely blown away by how moronic the vast majority of server-side code I've been exposed to is. Absurdly deep inheritance schemes. Getters/setters everywhere. Design patterns that make no sense in the context they've been placed in. How the hell is OOP any different from crap-procedural code when you have 25 methods from 25 different classes and objects all mutating the same data? It's just function spaghetti with redundant class syntax.
Type systems are about design tradeoffs (performance and compiler debug convenience vs. flexibity), not whether the language is closer to the one-true-way to do it right. Anybody who believes static types makes their code more robust is lying to themselves. If static typing protects you from anything it's from learning how not to write illegible, unmaintainable shit code that can only be tended to by using IDE crutches to work stupid faster.
In C/C++ I wouldn't have it any other way than static. When I'm writing UI and higher-level architecture, I'll take my variable length args, first-class functions, closures, extreme object mutability (but still not lacking for encapsulation), because when you actually have a little bit of discipline and attention to detail, you can get a lot more done with a lot less code and that is always easier to maintain, reuse and modify in the long haul.
It doesn't matter what the language was. Lack of proper support for the DOM standard on IE's part was the problem, not JavaScript. The ECMA spec has actually been very evenly supported since IE5. Always behind on the latest of course but at least we didn't have to polyfill native language methods in to replace stupid proprietary MS shenanigans for the core language.
But the last version of Netscape Navigator was closer to full compliance with CSS2 and the DOM specs than IE8 was. It took those assholes 10 !@#$ing years to add table-display properties. We've been normalizing the entire event system for IE's sake for a decade too until IE9 came out.
All of this by the way is why JS ended up reigning supreme on the client-side web. Nothing in popular use normalizes or handles UI concerns as well as JS does.
I think somebody's figured out mind control. Half of us can't be THIS stupid.
It was a letdown for me too. I was thinking more "plasma rifle." Awwww MICRO-wave : P
Okay, that's pretty yucky, but also basically DWI speed traps. This is cops pulling people over for no reason so that a private contracting firm can make 7.9 million dollars conducting a survey that will be statistically suspect at best. Cops doing something other than their job to interrupt people's lives not even for a DWI fishing expedition but acting as corporate minions.
Oh my god. They really HAVE violated the sanctity of our bodily fluids.
Most of these comments remind me of the conversations between the Scandinavian members of Dethklok.
What do they swab cheeks for other than DNA just out of curiosity?
But what if it's just for a completely useless survey? Come on! 10 bucks for some spit!
All non-free software is a rip-off and a scam, and I am convinced we should be able to prove it.
I agree with the majority of your post but I've always had an issue when people take it to this extreme. Is it maybe reasonable for a game developer to not give anybody and everybody permission to distribute their single-player PC-only game whose design assets alone cost millions to create? Is it even responsible to distribute the source of a web app that handles confidential user information? Is anybody who works for a non-free software company collaborating in a scam?
The Free Software Foundation kicks ass but if all the software of the world was free (yes I mean not as in beer but there is no real distinction for certain types of apps) there'd be little motivation for a lot specific types of software to have been made at the level of quality we've enjoyed in the first place.
Software isn't always just code. It's often composed of intellectual properties and other concerns that can't be protected if distributed freely. Actors like getting royalties for having their likenesses spread all over creation. Artists don't like having their distinctive works appropriated without being credited.
And what is a web-site but an app that's being distributed from one source only?
Coming from a web background, the idea of building a non-Free app using Free tools doesn't bother me one wit as long as you give back by contributing useful tools you developed in the process or giving feedback/support on the ones you used.
IMO, free and non-free apps are just a part of the same ecosystem. An Office Suite that let you rapidly port data for use in its various apps would have been a loonnnng time coming originating as an open-source endeavor. But there was serious money to be made there, which is a great motivator to find a way to get multiple teams on very different projects to learn to play nice together. This is not something open source communities collaborating between projects are necessarily famous for.
But as an alternative, a free version of an office suite also happened a lot sooner because MS has been a total douchebag about its software by turning interoperability into handcuffs. As long as OO has any kind of marketshare MS can only take the douchebaggery so far and I'd argue that free software has already directly and indirectly pushed them to make some positive changes. Everybody wins. As free tools and communities mature, non-free vendors will be forced to respond in kind and I suspect we'll all continue to win even more.
So don't see it as an adversarial thing. It's a relationship. A really fucked up dysfunctional one but that happens to produce a decent kid mostly by accident.
Just because open source devs seems smarter, doesn't meant they're paid any better. Hell, I'll bet they're typically paid less. A consequence of actually enjoying the work and not having to work as late perhaps.
Any office having trouble with a docx file in 2013 without the use of MSO is backward. Haven't had a problem opening in Google docs in years. Pretty sure OO's had that covered for a while too.
Also, winmail.dat is 100% a Microsoft-manufactured problem. They translate certain types of attachments to a proprietary Outlook format for security concerns real or imagined on SEND rather than simply buffering with said Outlook shenanigans on receipt but not resend so nobody else has to put up with Outlook's slow-ass dinosaur app BS but Outlook users. To be fair to MS, I suspec the Microsoft TwoFace coin landed on the incompetent moron side rather than the one featuring the slightly less incompetent devious asshole that day or they'd be doing the winmail.dat thing with a lot more file formats.
But fortunately that problem's already been compensated for with this plugin for Thunderbird. Found it at the top of my first Google. It's been out for a year. Maybe you should let your IT guys know before they get locked into... oh... too late aren't I?
https://addons.mozilla.org/en-us/thunderbird/addon/lookout/reviews/
Well at least when your users start pissing off some Thunderbird-using company that rightly identifies emails from clueless Outlook users as the problem, they can pass on that link so they don't have to pay money for the privilege of becoming a part of the same problem too.
So are not-root-canals when compared to root canals.
As a web UI developer I can promise you there have DEFINITELY been no such shenanigans with Internet Explorer which has blown goats honestly at least since 2000 and probably since IE first actually had competition.
I gathered from the article that the whole point was more, "Don't try to justify it in terms of savings or you will fail the second you hit a snag and it seems cheaper/easier to just go back than retrain anybody or do some research into some alternative you need."
Having grown up in a time when personal computers were just starting to get popular, you start to realize that some people will never be more than Foundation trilogy-style technopriests due to a deep and abiding fear of things they don't understand that literally paralyzes them into monkey see, monkey do behavior where technology is concerned. They are taught to click that icon, then that file thing in the gray bar and then "open" and then that folder name, then that file. Change that folder name or move it into subcategory somewhere right in front of their eyes and holy cow there's a problem. Move the entire file menu into a pretty windows icon button and it's brown trousers time. I'm not sure if it never occurs to them that there's some kind of a generalized process or scheme that they're following that could actually help them if they understood it at just one level of abstraction deeper or if they're just not actually capable of understanding things that way.
I see less of it now but still run into it even in younger people who've been around tech all their lives.
It's kind of like people who never truly learn to read properly. Beyond basic consonant use, words to them are mostly a series of letters they associate with words they speak rather than there being any real pattern to how the sequences of letters actually inform the sounds most of the time. My last name for instance is similar in vowel/consonant construction to "hatter." I'd say about half the time when somebody has to speak it out loud they screw up their eyes and think and then what comes out is "hater." It makes my wife, who is a human Oxford-English dictionary (including pronunciation keys) batshit.
The only way this surveillance crap will ever end well is if everybody gets to look at everybody's data. If I can check on where my political leaders, government officials, cops, and corporate leadership is at any time and see what they've been doing online, I have no objection to them doing the same to me or anybody else. Until that day, it's a problem. Because then it's not about whether I have anything to hide but rather whether they'd like to be able to misbehave without any real risk of me even trying to do anything or say anything about it.
As an ex-employee I'm all over dogging Sears but to be fair, that happened 21 years ago.
At one point the website would require you to navigate 8+ slowly-loading, cookie-dropping pages after you'd pulled the trigger on wanting to buy something. I once witnessed the legal department getting its 2 cents in and forcing change on some logos that it didn't like the look of. Purely aesthetic concern. It is the single-most dysfunctional corporate culture I've ever encountered.
Internet. 100% internet. But building those properties using the Monty Python castle in a swamp strategy.
They never were and never will be in a position to compete with Amazon:
http://www.tmz.com/2009/08/20/the-perfect-grill-for-a-cannibal/
At the time, they were running the show with 100s of offshore devs whose programming literacy I would place on a scale of non to bordering on competent. No version control. I am dead !@#$ing serious.
My vote's on corporations being allowed to buy votes with impunity, when what they and their lobbyists minions really deserve is to be drawn and quartered. I may not be entirely objective on the issue but I think that may have something to do with years of frustration of watching joe public getting beaten with a carrot and told to eat the stick and liking it.
Uh... the NSA HAS been so inclined. Lots of protesters getting accused of terrorism in recent years. We've got three kids in Chicago that have been in jail since the NATO summit. It took several months for the actual charges to be brought against them. They can do that now. Throw you in jail with no charges and come up with 'em later. If they follow the usual pattern, they'll drop the charges in a year or two when absolutely everybody has stopped paying attention. I'm not some Occupy kid. The police reaction to the protests was downright spooky. That the NSA was involved in a domestic affair is deeply disturbing.