This sort of thing is just terrible for a commercial ISV ecosystem.
It means your cost of validating your software goes through the roof, as you have a huge number of versions to test against.
Even when the APIs stay the same, other things change. And if nothing changed, then what would be the point of having a release?
Alkill below covers the other side, which is that it would be death in the corporate IT world.
This sort of thing only works marginally well in a pure open source environment, where there are people maintaining the repositories and updating everything for each new release. And then only when you stick to software from the repository. As soon as you start needing commercial non-open source software, then every API and ABI change just kills.
Of course, the flip side is that MS gets very ambitious when they only release new versions every 4-5 years. And they try to shoot the moon with a totally new thing where everything is changed. This also causes quite a bit of pain.
Big corporations are what they are. They sell us cool stuff with one hand and tighten the screws on our freedoms with the other. We hate them every morning and love them every afternoon, and vice versa. This is part of living in the modern world: you take your yin with your yang and try to figure out how to do what's right the best you can. If you think it has to be all one way or the other, that's cool, share your opinions, but don't expect everyone else to think the same.
Let's see. Bootcamp installs at basically the push of a button. Grub? Nope, have to do your own partitioning, then manually configure grub, after reading up on it to figure out how it works.
Bootcamp also has the windows drivers for the hardware, so even if you do use grub, you're going to have to use the resources from Bootcamp.
So lets recap.... a simple push-button solution that 'just works' or a bunch of work, both to produce exactly the same outcome.
Is it any wonder why more people dont just use grub?
Why would your computer manufacturer want to 'keep you from abandoning Vi$ta'?
If you have an XP disc & license, then use it.
If what you're saying is that you bought a machine that the manufacturer doesnt provide XP drivers for, then it sounds like you have an issue with the manufacturer.
The Class Action Lawsuit was over misleading 'Vista Capable' stickers, not over other companies (hp, dell, etc) not providing drivers.
For future reference, the way to avoid this situation is to not buy consumer targeted computers, especially never at a Retail Store like BestBuy or the like.
Buy corporate class equipment from HP, Dell, or Lenovo. Then you'll have not only better/more-stable drivers, but for all the commonly used operating systems as well, including often a couple linux versions.
Windows Media Player 11. I think it takes 200MB just to load its user interface. Are you kidding me? wmplayer.exe right no on my vista business x64 machine has a private working set of 25MB.
Thats with it playing a playlist of ~800 songs, full screened running visualizations.
It has to run virtualization out of the box. Pretty close. Hyper-V will be part of the OOBE by the end of the year, and in the meantime there are a number of free and simple to download and install options from MS and others.
It has to allow for per process and per CPU throttling Yeah, done with WSRM. Had that for years.
It has to run real time back up Yeah, done with VSS. For years.
support dedicated inline encryption and security subsystems I have no idea what this means, can you be more specific?
It has to support 16x more RAM 16x more than what? Windows supports up to 128GB of RAM in x86 versions and 2TB of memory in x64 versions. Do you really require 32TB of memory to be functional?
and an order of magnitude larger AD spaces Order of magnitude larger than what? A single forest can reliably hold over 1B objects. If you need more than that, you use more than one forest.
It has to support virtualized patches If you mean binary differential patches, they've been using this for the best part of a decade.
a journalled file system Yeah, thats been there since day 1 of NTFS. So old in the windows world its archaic.
a file system that spans physical volumes. Yeah, dynamic volumes has been able to do that for like 6+ years.
The word 'evil' isnt even relevant in the realm of the software industry.
Evil people cause human beings to die, or to be tortured, injured, etc. Evil is not a word to describe higly competitive (and highly effective) business execution.
It's just business man, no one has an obligation to make any given business succeed. At the worst, the founders/owners/investors may lose some time or money. But you know what? Thats life for a business owner. It's risky. Thats life.
To those losing their jobs as their companies get killed by anticompetetive practices it doesn't really help that much (obviously) that there were nice people at Microsoft. Give me a break. It's a business. Sometimes people lose jobs. Sometimes one company destroys another by being so much more successful that the other business is crowded out. This is regular life in the business world.
To those getting downsized as corporate expenditure on software made the business unprofitable it doesn't help either. If any business has to downsize staff because they're spending so much money on commercial software, and that makes them non-competitive in their market, then they're going to fail, and they deserve to. It means that the owners dont know how to run a business properly. And in any case, people get 'downsized' for much stupider reasons.
I'd suggest that the nice guys in the trenches, if they want to demonstrate their good will, ensure that the evil guys are the first one over the top the next time. And that they get shot in the back repeatedly just in case the enemy doesn't do the job right. Again, you seem to be confusing real harm like people dying (you speak of wars and trenches) with business, and jobs. The two are not the same. Business is not war. At worst, in business, people lose money. In war, people lose their lives.
If you cant tell the difference between the two, then you probably shouldnt be posting about it.
It's turtles all the way down. In Smalltalk, everything is an object, and all operations are defined by message sends. This even extends to flow control. In Smalltalk, there is a Boolean class with two concrete (singleton) subclasses, True and Fales. Boolean objects respond to an ifTrue: message with a block (closure) as an argument. Instances of the True class execute the argument, instances of False do not.... Because all flow control is defined in terms of objects, you can modify them or add your own which are indistinguishable from those provided by the language. Even classes are defined in the same way; to create a new class you instantiate a metaclass and then send the resulting instance messages which add methods and instance variables. It seems like this would be a source of problems. If classes can be modified at run-time, how could an IDE ever know what all the methods available on a class are? If you wanted the IDE to tell you where all calls against a certain method are, and what class they're being called against, and the call stack there, etc, how would this be done?
It seems like you can never know what classes will be like, because they could change conditionally, from one run to the next.
You can modify the UI while the program runs in this way whenever you find something wrong with it. Because there is no separation between the debugging and coding phases, you fix bugs as soon as their symptoms manifest, which is a much more robust approach. This stuff is huge. Now if only I could get that with strongly typed languages and proper interfaces (as opposed to duck typing, which seems to me like a tiny implicit interface for every method), I'd be very happy. I miss not having to go through compile/deploy cycles. Just modify and run.
Good point, I've heard this as well, but havent yet made the time to do some personal projects in smalltalk.
I would love to know what the Smalltalk folks are doing that the Ruby/Python folks arent. Is it something fundamentally different in the language? For example, doesnt Smalltalk support optionally strongly typing members?
I've never had a good understanding of why Smalltalk can do this but the other modern (though much younger) dynamic languages (python & ruby particularly) cant do it.
It's also had debugging features since the mid '70s that make modern IDEs look incredibly primitive, but that's a topic for another discussion. I'm skeptical of this statement, compared to current systems. Maybe up through the mid-90's, but modern debuggers are pretty impressive. I'm not sure what else you could add. You can break on method names, on exception types, on modifications to member and memory addresses. Some modern languages and stop-and-continue where you can modify the code while in break-mode in the debugger, and then just continue.
What features do you see in the Smalltalk debugger that you dont see in other modern debuggers?
You dont understand code signing. No one 'approves' your executable. They dont even ever SEE your executable.
All the code signing does is create a followable trail back to an actual person or corp. It's non-repudiation for software.
When you buy a code signing cert from Verisign or similar, you get just that, a code signing cert with your company name attached to it. Then YOU sign your executables and distribute them. No one else ever sees or approves your code.
And it costs ~$200 per year, if you renew it every year.
And if you want to run untrusted code, then you just run untrusted code.
In addition, there are hobbyist toolkits for at least the XBox 360 (not sure about PS3 or Wii). The software is free (iirc) and there's like a $20 'club' you need to join to be able to load your home-developed code on the xbox.
Are the code completion and jump-to-declaration actually based off syntax trees or just text searching? There's a fundamental different in behavior between the two. The latter is quite hacky compared to the former. Of course, the former only works right on a strongly-typed 'compiled' language.
I don't see the point in code completion. If I don't know what some class is doing, I'd much rather read the documentation than just browsing through the list of methods. It might save a few seconds once in a while, but that's about it. The thing is, you only need to read the documentation once ever usually, but just because you've read it doesnt mean you remember every little detail about the various overloaded methods, what parameters, or even the spelling on longer methods, or between many similarly named methods across large libraries (StringUtils.isBlank() vs. StringUtils.isEmpty(), for example).
In addition, what's easier and faster (and therefore less likely to break you out of the zone):
- type the first couple characters, then CTRL+SPACE, and you see all the matching methods PLUS their javadocs. So you have not only code completion but the documentation right there in the IDE. No need to go out to a browser or PDF.
- type the whole thing exactly right, every time, even on libraries you only use once or twice a year. And when you dont remember, you have to go out to a browser, load up the documentation, search for it, and then read it, then go back to the IDE or text editor.
But if you're doing 'space qualified software' you're probably working with much smaller libraries, and a much more tightly constrained scope, so you may never see the difference. Typical corporate development, on the other hand, you not only have all the built-ins of the language, but the huge class libraries (.net and java) that ship with them, but then the entire project you're working on. When just the project can consist of hundreds of classes and many thousands of methods, there's just no way I can remember them all, in addition to having memorized the entire.net/java class libraries. If you can, I'm impressed.
I cant say off the top of my head (dont feel like firing it up) if VS 2005 or 2008 has it, but Eclipse has the 'back' function that behaves just like you'd expect (and what you're describing).
Thats interesting, I didnt know that the Mozilla folks came up with the now-ubiquitous RSS feed icon. Thats pretty cool.
And FireFox's dragon (or whatever) icon for the main product is cool. It's the rest of the browser that just looks amateurish, at least IMO.
Wonder if they have different folks doing the icons than the rest of the browser UI.
I also have a memory of news that Mozilla was finally going to hire professional graphic designers for their UI. Cant find a link offhand though, so could be mis-remembering.
Now that was just a ridiculous response to a very reasonable statement by the parent.
Code completion is a crutch and not particularly helpful: "oooh look a menu of stuff to scroll through and select the right function!"... now honestly, was that any faster than just typing the damn function? You're purposefully being dense. Java or.NET have a huge number of APIs. No one memorizes every single one, and every single overload of every API, and every single parameter type and description for every single one. Thats not reasonable.
Why not have the computer do what the computer is good at, and remember/index things? Why would you clog your brains up with trying to memorize that stuff, or slow yourself down by having to swap out to javadocs or help files?
You have intellisense/code-completion so that you dont have to type every little function, and so you dont have to remember the exact spelling and parameter ordering of every single function.
Heck, even medium sized projects I've worked on have > 1000 classes, with several methods each. No way is anyone going to memorize every single one of these, plus the javadocs, etc.
As for project-wide renames, learn sed, my friend. This statement just shows you dont get it. SED relies on text matching, which is imperfect (picks up on comments, javadocs, or methods that contain the method name you want as a substring). Refactoring (ie, project-wide renames) use a syntax tree of the actual calls. So it knows precisely where every single call of a specific method is, and doesnt rely on brute force and unreliable text searching and pattern matching.
You can do some amazing things with a syntax tree of your software. sed/grep and the like are powerful tools for what they're for, but dont compare to what true refactoring tools in a strongly typed language can do.
With one button hit (one stroke of a finger) I can go to the declaration of a method. No waiting, no searching, just instantly goes there. With another 1-stroke, I can go to a list of all places in the code where that method is called. No waiting, no searching, no results that include comments or text or xml files. And then from there you can brows up or down the call stack.
The refactoring tools in Eclipse are superb. Better than what VS offers without additional purchased software.
And oh how I wish I could have refactoring, intellisense, code completion, and all the beautiful things Eclipse gives me in that area for a dynamic language like Ruby. One can dream I guess.
Maybe Python or Ruby will offer an optional explicit typed mode. Again, one can dream.
You could just make a personal branch (or personal project, but that will be less elegant).
It's not as simple as the TFS shelving idea, but it works, and isnt too bad.
Fortunately, losing a day or two's worth of work can almost always be recovered in half that time or less. Unless you're in a death march, or it happens all the time, losing a day or two isnt a huge showstopper.
First, although you are targeting in on Opera, I was not. I know nothing of Opera's internal business organization, or whether my description of how it can work at a generic large org matches what Opera looks like internally. Nor was I suggesting that Opera's response this time was bad or good.
Neither of us know anything abou Opera's business organization, so you probably shoudnt be so quick to jump to judgement about something you may not know anything about. This article, in particular, was short on cold hard facts and high on the 'read between the lines'. So very few of us posting here know what actually happened.
There should be a predefined process for fast escalation; As soon as it hits the level of "this might be a real bug", a developer should be looking into it. As I was trying to state, its not always that easy. And the devil in the details is often making the decision that 'this might be a real bug'. Thats not always a trivial step.
And think about how that might work in the real world. You've found a big hole in a private company (ie, not open source). It's an own your box situation, but its complicated to demonstrate, and is dependent on several things being present, platforms, etc.
How do you tell 'Opera' (or another similar company) that there is a bug, and that its severe?
For the company side of things, who want to know this information, they're faced with a conundrum. Make it real easy to contact them about these bugs, and have a very low signal to noise ratio (ie, alot of useless emails, general ranting, complaints about the product). Reality is that if you have a contact form on your website, even if there's big warning labels all over it that its only for severe bug reporting, it'll get spammed to obvlivion. People will use it to bitch at the company, make complaints or suggestions about the product, or give useless bug reports (Opera crashed... can you fix it?).
It takes real human beings with real experienced judgement ability to separate the signal from the noise.
Or a company can make it hard to contact them with this information, which means only quality stuff gets to them, but it can be hard for the newb to pass the information up.
It can get even worse! The bug reporter may have just the right person, and have a glorious repro demo steps outlined, but they dont speak very well, so the contact comes across as amateurish and useless. That happens. Alot.
Do you see the problem? Just having people 'standing by' and ready is not very useful. You need to have a way for the people with the real issues to get heard, but while making the noise level low enough to be functional in. This is not easy.
And the reality is that many companies do this on an ad-hoc person-to-person basis. Some guy knows some other guy, who knows a guy at the company. These are not structures that lend themselves to fast responses.
Even the open source companies struggle with this. Go look at FireFox's bug list. How many hundreds of people report certain types of problems and the developers just basically give them the finger, because they feel they've heard it all before. Or its not repro'able. Or its hard to understand. Or it doesnt work on their machine.
You make it appear from your post that someone can just declare, 'Make it so!' and it'll work perfectly. It wont. The best of companies struggle with this. The best of open source projects struggle with this. It is a fundamentally 'hard' problem in CS speak.
And unfortunately, one of the most effective ways to bypass these networks of people is to make it public and raise some stink. Then the company is sure to see it and respond. But its not a very nice way to get things done.
Anyway, my point is that its not as easy as you make it sound. Especially in the case of companies whose products sell or distribute to the general public.
I do wish Opera would take this update opportunity to fix their toolbar so it looks similar to IE and Firefox, in that the blank space, where Opera used to have their advertisement bar, is removed, and filled with browser controls like the others have. To me, the greatest thing is Firefox having the toolbar editor, so the user can set it up like they want. Do you realize that Opera's entire GUI is completely user-configurable, without any plugins?
You just right click on the toolbar, click Customize, then drag and drop to your heart's content. Couldnt be easier.
I'm not sure what blank space you're talking about. My Opera (on windows) have no blank space. And even if it did, you just re-organize the toolbars to eliminate it.
Heck, you can even put the tabs (or any toolbar or menu bar) on the side of the screen or the bottom (where I prefer) if you want.
In my opinion, Opera has a much cleaner toolbar than either Firefox (very amateur, blocky) or IE (schizophrenic, why are half the buttons on one side, half on the other?).
Firefox's GUI in particular always looks very amateurish. Like it was done by 'this guy' that someone knew who 'is good with graphics'. Whereas the other browsers actually hired professionals.
The problem usually isnt coding time. It's organizational response and resource allocation issues.
For example, Opera is on a very differen timezone from the US, so initial publication may happen overnight from the POV of the Opera staff.
So then a day starts. When people start their day, they have a pile of things to respond to. The incoming messsages have to be triaged. Someone has to make a decision that this is important enough to escalate or take action on.
Then you have to find people with the capability to test whether its a real problem. This may take a couple hours. People go on vacation, get sick, etc.
Then you have to take the time to do the research, test whether this is a real problem, what versions it affects, etc. This takes a couple hours.
Then yuou have to stop a coder from working on something else, bring them up to speed on the problem (if its not the same person doing the testing), and get them started on the fix.
Then even with a fix you have to do regression tests. Not sure about Opera, but many mature apps have full test suites that can take a couple hours.
Then you have to write release notes, update the web page, do a new deploy package, and update your update servers to notify Opera that there is a new update.
As you can see, very little of the time here is coding.
Many large orgs have taken steps to create a 'short path of decision making' to streamline this process, always have one coder on call who can do this work, etc. But even then if anything is out of whack or the wrong person is sick or on vacation or on another urgent item, a whole day could pass without response.
Yeah, lovely. Just what I want to do is sit in a grease-shack and work in a hard plastic chair smelling french fries, while sharing the space with homeless people (speaking to the fast food joints offering wifi).
I'm sure they're making a good profit, nothing wrong with that if they can convince people to pay it.
But its not like all that money is flowing right back to shareholders and C-level executives. Starbucks pays their baristas well (compared to other coffee shops), they pay full benefits, 401k, stock options, and frankly makes for an excellent place to work for those in the position to want that kind of a job.
So their expenses are going to be higher than your typical mom & pop store, which, all things being equal, means their prices will also be higher. Of course, their supply chain & distribution network is probably far more efficient than a mom&pop, so not sure where it balances out.
I guess my biggest problem with your posts though are the unnecessary loaded words. Things like 'exploit' and 'know no better' which brings out imagery of greedy evil geniuses sitting around a republican board room figuring out how next to bilk the stupid masses.
But believe it or not, there are those of us who actually like and prefer Starbucks. And do so through an informed, educated choice.
I've tried ALL the coffee shops where I live. For most of them, the regular coffee (drip) is weak and tasteless. And their fancy drinks are also not very good. And they look at you weird and do a crappy job if you want something simple like 4 shots of espresso over ice (for those places who dont do iced toddy (sp?)). And several of the locals have horribly inconsistent performance. It's a gamble from one day to the next what quality of coffee you get.
And the food is crappier, and the baristas are not as friendly, and the wifi is slower.
This doesnt even get into the fact that with Starbucks & T-Mobile, I can go pretty much anywhere in the US (where I live) and get exactly the coffee I want and like, know exactly how it'll taste, wont get any hassle from the baristas (and get an incredibly nice smiling face), and have a fast, reliable internet connection. This is true pretty much everywhere in the US, even down in the depths of Indiana and equivalent places.
Now mind you I've run into some equally good places I like. Caribou Coffee being one, but there isnt one where I live. I also have been to some amazing local coffee shops in Seattle, and I keep hearing about great ones in the bay area. But locals that are that good are few and far between.
Slashdot Mirror
This sort of thing is just terrible for a commercial ISV ecosystem.
It means your cost of validating your software goes through the roof, as you have a huge number of versions to test against.
Even when the APIs stay the same, other things change. And if nothing changed, then what would be the point of having a release?
Alkill below covers the other side, which is that it would be death in the corporate IT world.
This sort of thing only works marginally well in a pure open source environment, where there are people maintaining the repositories and updating everything for each new release. And then only when you stick to software from the repository. As soon as you start needing commercial non-open source software, then every API and ABI change just kills.
Of course, the flip side is that MS gets very ambitious when they only release new versions every 4-5 years. And they try to shoot the moon with a totally new thing where everything is changed. This also causes quite a bit of pain.
The tagline is 'News for Nerds, Stuff that Matters'. Nothing about Free and Open Source there.
The closest thing that I can find is in the FAQ:
http://slashdot.org/faq/slashmeta.shtml#sm1100 I thought everyone on Slashdot hated the RIAA, the MPAA, and Microsoft. Why do you keep hyping CDs, movies, and Windows games?
Big corporations are what they are. They sell us cool stuff with one hand and tighten the screws on our freedoms with the other. We hate them every morning and love them every afternoon, and vice versa. This is part of living in the modern world: you take your yin with your yang and try to figure out how to do what's right the best you can. If you think it has to be all one way or the other, that's cool, share your opinions, but don't expect everyone else to think the same.
Why not just use grub?
.... a simple push-button solution that 'just works' or a bunch of work, both to produce exactly the same outcome.
Let's see. Bootcamp installs at basically the push of a button. Grub? Nope, have to do your own partitioning, then manually configure grub, after reading up on it to figure out how it works.
Bootcamp also has the windows drivers for the hardware, so even if you do use grub, you're going to have to use the resources from Bootcamp.
So lets recap
Is it any wonder why more people dont just use grub?
Why would your computer manufacturer want to 'keep you from abandoning Vi$ta'?
If you have an XP disc & license, then use it.
If what you're saying is that you bought a machine that the manufacturer doesnt provide XP drivers for, then it sounds like you have an issue with the manufacturer.
The Class Action Lawsuit was over misleading 'Vista Capable' stickers, not over other companies (hp, dell, etc) not providing drivers.
For future reference, the way to avoid this situation is to not buy consumer targeted computers, especially never at a Retail Store like BestBuy or the like.
Buy corporate class equipment from HP, Dell, or Lenovo. Then you'll have not only better/more-stable drivers, but for all the commonly used operating systems as well, including often a couple linux versions.
Thats with it playing a playlist of ~800 songs, full screened running visualizations.
The word 'evil' isnt even relevant in the realm of the software industry.
Evil people cause human beings to die, or to be tortured, injured, etc. Evil is not a word to describe higly competitive (and highly effective) business execution.
It's just business man, no one has an obligation to make any given business succeed. At the worst, the founders/owners/investors may lose some time or money. But you know what? Thats life for a business owner. It's risky. Thats life. To those losing their jobs as their companies get killed by anticompetetive practices it doesn't really help that much (obviously) that there were nice people at Microsoft. Give me a break. It's a business. Sometimes people lose jobs. Sometimes one company destroys another by being so much more successful that the other business is crowded out. This is regular life in the business world. To those getting downsized as corporate expenditure on software made the business unprofitable it doesn't help either. If any business has to downsize staff because they're spending so much money on commercial software, and that makes them non-competitive in their market, then they're going to fail, and they deserve to. It means that the owners dont know how to run a business properly. And in any case, people get 'downsized' for much stupider reasons. I'd suggest that the nice guys in the trenches, if they want to demonstrate their good will, ensure that the evil guys are the first one over the top the next time. And that they get shot in the back repeatedly just in case the enemy doesn't do the job right. Again, you seem to be confusing real harm like people dying (you speak of wars and trenches) with business, and jobs. The two are not the same. Business is not war. At worst, in business, people lose money. In war, people lose their lives.
If you cant tell the difference between the two, then you probably shouldnt be posting about it.
Cool, thanks for the information. One more language I need to try out.
Because all flow control is defined in terms of objects, you can modify them or add your own which are indistinguishable from those provided by the language. Even classes are defined in the same way; to create a new class you instantiate a metaclass and then send the resulting instance messages which add methods and instance variables. It seems like this would be a source of problems. If classes can be modified at run-time, how could an IDE ever know what all the methods available on a class are? If you wanted the IDE to tell you where all calls against a certain method are, and what class they're being called against, and the call stack there, etc, how would this be done?
It seems like you can never know what classes will be like, because they could change conditionally, from one run to the next. You can modify the UI while the program runs in this way whenever you find something wrong with it. Because there is no separation between the debugging and coding phases, you fix bugs as soon as their symptoms manifest, which is a much more robust approach. This stuff is huge. Now if only I could get that with strongly typed languages and proper interfaces (as opposed to duck typing, which seems to me like a tiny implicit interface for every method), I'd be very happy. I miss not having to go through compile/deploy cycles. Just modify and run.
I dont know how you get that from my post.
... then just dont do it.
I was correcting the parent AC's misunderstanding of what code signing is and how its used.
But if you dont like that sort of thing
As a developer, dont sign your code. As an end-user, dont not run code just because its unsigned.
I would love to know what the Smalltalk folks are doing that the Ruby/Python folks arent. Is it something fundamentally different in the language? For example, doesnt Smalltalk support optionally strongly typing members?
I've never had a good understanding of why Smalltalk can do this but the other modern (though much younger) dynamic languages (python & ruby particularly) cant do it. It's also had debugging features since the mid '70s that make modern IDEs look incredibly primitive, but that's a topic for another discussion. I'm skeptical of this statement, compared to current systems. Maybe up through the mid-90's, but modern debuggers are pretty impressive. I'm not sure what else you could add. You can break on method names, on exception types, on modifications to member and memory addresses. Some modern languages and stop-and-continue where you can modify the code while in break-mode in the debugger, and then just continue.
What features do you see in the Smalltalk debugger that you dont see in other modern debuggers?
You dont understand code signing. No one 'approves' your executable. They dont even ever SEE your executable.
All the code signing does is create a followable trail back to an actual person or corp. It's non-repudiation for software.
When you buy a code signing cert from Verisign or similar, you get just that, a code signing cert with your company name attached to it. Then YOU sign your executables and distribute them. No one else ever sees or approves your code.
And it costs ~$200 per year, if you renew it every year.
And if you want to run untrusted code, then you just run untrusted code.
In addition, there are hobbyist toolkits for at least the XBox 360 (not sure about PS3 or Wii). The software is free (iirc) and there's like a $20 'club' you need to join to be able to load your home-developed code on the xbox.
In addition, what's easier and faster (and therefore less likely to break you out of the zone):
- type the first couple characters, then CTRL+SPACE, and you see all the matching methods PLUS their javadocs. So you have not only code completion but the documentation right there in the IDE. No need to go out to a browser or PDF.
- type the whole thing exactly right, every time, even on libraries you only use once or twice a year. And when you dont remember, you have to go out to a browser, load up the documentation, search for it, and then read it, then go back to the IDE or text editor.
But if you're doing 'space qualified software' you're probably working with much smaller libraries, and a much more tightly constrained scope, so you may never see the difference. Typical corporate development, on the other hand, you not only have all the built-ins of the language, but the huge class libraries (.net and java) that ship with them, but then the entire project you're working on. When just the project can consist of hundreds of classes and many thousands of methods, there's just no way I can remember them all, in addition to having memorized the entire
I cant say off the top of my head (dont feel like firing it up) if VS 2005 or 2008 has it, but Eclipse has the 'back' function that behaves just like you'd expect (and what you're describing).
I use it in the ALT+LEFT-ARROW form all the time.
Dont normally like feeding the cowards, but ..
Thats interesting, I didnt know that the Mozilla folks came up with the now-ubiquitous RSS feed icon. Thats pretty cool.
And FireFox's dragon (or whatever) icon for the main product is cool. It's the rest of the browser that just looks amateurish, at least IMO.
Wonder if they have different folks doing the icons than the rest of the browser UI.
I also have a memory of news that Mozilla was finally going to hire professional graphic designers for their UI. Cant find a link offhand though, so could be mis-remembering.
Why not have the computer do what the computer is good at, and remember/index things? Why would you clog your brains up with trying to memorize that stuff, or slow yourself down by having to swap out to javadocs or help files?
You have intellisense/code-completion so that you dont have to type every little function, and so you dont have to remember the exact spelling and parameter ordering of every single function.
Heck, even medium sized projects I've worked on have > 1000 classes, with several methods each. No way is anyone going to memorize every single one of these, plus the javadocs, etc. As for project-wide renames, learn sed, my friend. This statement just shows you dont get it. SED relies on text matching, which is imperfect (picks up on comments, javadocs, or methods that contain the method name you want as a substring). Refactoring (ie, project-wide renames) use a syntax tree of the actual calls. So it knows precisely where every single call of a specific method is, and doesnt rely on brute force and unreliable text searching and pattern matching.
You can do some amazing things with a syntax tree of your software. sed/grep and the like are powerful tools for what they're for, but dont compare to what true refactoring tools in a strongly typed language can do.
With one button hit (one stroke of a finger) I can go to the declaration of a method. No waiting, no searching, just instantly goes there. With another 1-stroke, I can go to a list of all places in the code where that method is called. No waiting, no searching, no results that include comments or text or xml files. And then from there you can brows up or down the call stack.
The refactoring tools in Eclipse are superb. Better than what VS offers without additional purchased software.
And oh how I wish I could have refactoring, intellisense, code completion, and all the beautiful things Eclipse gives me in that area for a dynamic language like Ruby. One can dream I guess.
Maybe Python or Ruby will offer an optional explicit typed mode. Again, one can dream.
You could just make a personal branch (or personal project, but that will be less elegant).
It's not as simple as the TFS shelving idea, but it works, and isnt too bad.
Fortunately, losing a day or two's worth of work can almost always be recovered in half that time or less. Unless you're in a death march, or it happens all the time, losing a day or two isnt a huge showstopper.
First, although you are targeting in on Opera, I was not. I know nothing of Opera's internal business organization, or whether my description of how it can work at a generic large org matches what Opera looks like internally. Nor was I suggesting that Opera's response this time was bad or good.
Neither of us know anything abou Opera's business organization, so you probably shoudnt be so quick to jump to judgement about something you may not know anything about. This article, in particular, was short on cold hard facts and high on the 'read between the lines'. So very few of us posting here know what actually happened. There should be a predefined process for fast escalation; As soon as it hits the level of "this might be a real bug", a developer should be looking into it. As I was trying to state, its not always that easy. And the devil in the details is often making the decision that 'this might be a real bug'. Thats not always a trivial step.
And think about how that might work in the real world. You've found a big hole in a private company (ie, not open source). It's an own your box situation, but its complicated to demonstrate, and is dependent on several things being present, platforms, etc.
How do you tell 'Opera' (or another similar company) that there is a bug, and that its severe?
For the company side of things, who want to know this information, they're faced with a conundrum. Make it real easy to contact them about these bugs, and have a very low signal to noise ratio (ie, alot of useless emails, general ranting, complaints about the product). Reality is that if you have a contact form on your website, even if there's big warning labels all over it that its only for severe bug reporting, it'll get spammed to obvlivion. People will use it to bitch at the company, make complaints or suggestions about the product, or give useless bug reports (Opera crashed
It takes real human beings with real experienced judgement ability to separate the signal from the noise.
Or a company can make it hard to contact them with this information, which means only quality stuff gets to them, but it can be hard for the newb to pass the information up.
It can get even worse! The bug reporter may have just the right person, and have a glorious repro demo steps outlined, but they dont speak very well, so the contact comes across as amateurish and useless. That happens. Alot.
Do you see the problem? Just having people 'standing by' and ready is not very useful. You need to have a way for the people with the real issues to get heard, but while making the noise level low enough to be functional in. This is not easy.
And the reality is that many companies do this on an ad-hoc person-to-person basis. Some guy knows some other guy, who knows a guy at the company. These are not structures that lend themselves to fast responses.
Even the open source companies struggle with this. Go look at FireFox's bug list. How many hundreds of people report certain types of problems and the developers just basically give them the finger, because they feel they've heard it all before. Or its not repro'able. Or its hard to understand. Or it doesnt work on their machine.
You make it appear from your post that someone can just declare, 'Make it so!' and it'll work perfectly. It wont. The best of companies struggle with this. The best of open source projects struggle with this. It is a fundamentally 'hard' problem in CS speak.
And unfortunately, one of the most effective ways to bypass these networks of people is to make it public and raise some stink. Then the company is sure to see it and respond. But its not a very nice way to get things done.
Anyway, my point is that its not as easy as you make it sound. Especially in the case of companies whose products sell or distribute to the general public.
You just right click on the toolbar, click Customize, then drag and drop to your heart's content. Couldnt be easier.
I'm not sure what blank space you're talking about. My Opera (on windows) have no blank space. And even if it did, you just re-organize the toolbars to eliminate it.
Heck, you can even put the tabs (or any toolbar or menu bar) on the side of the screen or the bottom (where I prefer) if you want.
In my opinion, Opera has a much cleaner toolbar than either Firefox (very amateur, blocky) or IE (schizophrenic, why are half the buttons on one side, half on the other?).
Firefox's GUI in particular always looks very amateurish. Like it was done by 'this guy' that someone knew who 'is good with graphics'. Whereas the other browsers actually hired professionals.
The problem usually isnt coding time. It's organizational response and resource allocation issues.
For example, Opera is on a very differen timezone from the US, so initial publication may happen overnight from the POV of the Opera staff.
So then a day starts. When people start their day, they have a pile of things to respond to. The incoming messsages have to be triaged. Someone has to make a decision that this is important enough to escalate or take action on.
Then you have to find people with the capability to test whether its a real problem. This may take a couple hours. People go on vacation, get sick, etc.
Then you have to take the time to do the research, test whether this is a real problem, what versions it affects, etc. This takes a couple hours.
Then yuou have to stop a coder from working on something else, bring them up to speed on the problem (if its not the same person doing the testing), and get them started on the fix.
Then even with a fix you have to do regression tests. Not sure about Opera, but many mature apps have full test suites that can take a couple hours.
Then you have to write release notes, update the web page, do a new deploy package, and update your update servers to notify Opera that there is a new update.
As you can see, very little of the time here is coding.
Many large orgs have taken steps to create a 'short path of decision making' to streamline this process, always have one coder on call who can do this work, etc. But even then if anything is out of whack or the wrong person is sick or on vacation or on another urgent item, a whole day could pass without response.
Dont forget about the $100+ per month forever for commercial dsl to serve the wifi.
Yeah, lovely. Just what I want to do is sit in a grease-shack and work in a hard plastic chair smelling french fries, while sharing the space with homeless people (speaking to the fast food joints offering wifi).
I'm sure they're making a good profit, nothing wrong with that if they can convince people to pay it.
But its not like all that money is flowing right back to shareholders and C-level executives. Starbucks pays their baristas well (compared to other coffee shops), they pay full benefits, 401k, stock options, and frankly makes for an excellent place to work for those in the position to want that kind of a job.
So their expenses are going to be higher than your typical mom & pop store, which, all things being equal, means their prices will also be higher. Of course, their supply chain & distribution network is probably far more efficient than a mom&pop, so not sure where it balances out.
I guess my biggest problem with your posts though are the unnecessary loaded words. Things like 'exploit' and 'know no better' which brings out imagery of greedy evil geniuses sitting around a republican board room figuring out how next to bilk the stupid masses.
But believe it or not, there are those of us who actually like and prefer Starbucks. And do so through an informed, educated choice.
I've tried ALL the coffee shops where I live. For most of them, the regular coffee (drip) is weak and tasteless. And their fancy drinks are also not very good. And they look at you weird and do a crappy job if you want something simple like 4 shots of espresso over ice (for those places who dont do iced toddy (sp?)). And several of the locals have horribly inconsistent performance. It's a gamble from one day to the next what quality of coffee you get.
And the food is crappier, and the baristas are not as friendly, and the wifi is slower.
This doesnt even get into the fact that with Starbucks & T-Mobile, I can go pretty much anywhere in the US (where I live) and get exactly the coffee I want and like, know exactly how it'll taste, wont get any hassle from the baristas (and get an incredibly nice smiling face), and have a fast, reliable internet connection. This is true pretty much everywhere in the US, even down in the depths of Indiana and equivalent places.
Now mind you I've run into some equally good places I like. Caribou Coffee being one, but there isnt one where I live. I also have been to some amazing local coffee shops in Seattle, and I keep hearing about great ones in the bay area. But locals that are that good are few and far between.