Which is why I think there should be a pay floor of 125K/year on H1-B visa workers.. if there's nobody here to fill that job, it must require someone special, which means the pay should be that much higher or more.
Well, an experienced dev is likely to have 3-10x the output with a fraction of the bugs or other rework. Even the math works in the favor of a more experienced dev.
I usually spend the first couple hours 2-3 days a week reading blog posts, and keeping up... there are a huge amount of things to play with, choices seem to double every year, and many trends or toy frameworks don't make it. It's impossible to know everything in depth.. but it is entirely possible to spot trends, and learn what is becoming more common place. Right now, if you aren't at least looking at Node.js or Python, and contemplating a NoSQL database, you are going to be behind in a few years. Java and C#.Net are tried and true, and will be around for a while. Hell C/C++ and COBOL are still really broadly used. They're all needed and serve a purpose.
Ironically I think what was old is new again... That scripting environments and various data store options are coming to the front again. Scripting runtimes, and computing power are better than ever, and SQL doesn't fit every need well. I think in another 10 years there will be new dev environments... One size doesn't fit all... There is no perfect hammer... but what we can look for are better fits for a given need.
My best suggestion to a younger developer is don't get locked into one thing... think about the problems you are tasked to resolve and know your chosen environment as well as possible to resolve those problems. Keep your eyes open to new things... try a new language every year.. there's usually a couple to choose from. You don't have to do anything meaningful, or in the workplace to glean knowledge from something new. If something doesn't fit, don't force it.. and if something works better, push for it... don't get caught up in politics, and choose wisely.
I've seen it go both ways.. I'm pushing 40, and do a pretty good job keeping up, and being aware of new ideas, concepts and tools... on the flip side, I had a coworker that was a couple years older absolutely resistant to any new tools, techniques or change in general... Then again, I've seen plenty of younger developers who are relatively fresh from college who can't think their way through, under or around a given problem.
I've found that more and more a kanban approach tends to make people a little happier... tasks come in, tasks go out... the next task is prioritized, and interaction with others during a coding session should be mitigated by your manager... I'm also finding a unix-like approach of small bits of software that can pump in/out logic separately are a bit better than monolithic tightly coupled libraries... less OO, and more functional piped abstractions. Today I'm doing more of this in Node.js, it lets me get stuff done, usually with less code, tooling and overhead. You could do similar in C/C++, but that brings a compilation step into the mix. When I have to fire up an IDE and do a build to check/test/run/change anything, it just seems like overhead that shouldn't be needed anymore.
Using unix-like principles, and node (or any other scripted language) lets me get stuff done... following conventions make it more maintainable. I'm 37 and still learning... I can imagine I'll still be doing so for another few decades.
True enough.. I had both a Chevy Silverado 1500 crew cab, and a Hyundai Sonata before my current car a Challenger RT... I tended to do about 5-6mpg higher than rated in the sonata, and about the rating in the silverado... about 2/3 my driving was highway at the time... in my challenger, it's a bigger difference.. I tend to do best around 75mph relatively level than my current drive, about 1/2 highway, and heavy traffic... my general driving is around 19-20mpg, but I made a fairly decent trip a few weeks back averaging around 70-80mph, and got nearly 28mpg. How you drive has a big impact.. but traffic kills any predictability with fuel economy.
I don't load most apps that require internet access, or for that matter contacts etc... there are few exceptions, but those are for apps that are related to the permissions they are asking for. I'm doubly leery when an app doesn't auto update, usually indicates a permission change.
given that my actual bank password cannot be anything other than us letters and numbers, no special characters (at two banking institutions)... An incident where I had over a million dollars in my account (magically) for a couple days, another where a bank error cost me a few hundred, and another still when a merger lost a friend over 10K with no tracking ability... I don't trust banks all that much either.
That doesn't make it a bad decision to stick with it... I find that more recent versions of gnome/unity/kde can be very sluggish on even modern hardware. I just want a usable desktop with a modest amount of eye candy that runs well. I really like the Win7 taskbar the best currently... Would say that the Mint desktop is probably a distant second, followed by osx. It's sad MS had to f*ck it all up with Win8.
As much as I don't even care for the surveillance aspects... It's highly unlikely that local law enforcement will have weaponized drones any time soon... In much the same way as they don't typically carry RPGs.
That isn't nearly so true... If you've used many ARM devices, there are a lot of projects that take a little, to a lot of tweaking for a non x86 environment.
Why have gui programs on your servers to begin with? I keep hearing this touted as the biggest reason for keeping X over Wayland, but still fail to see WTF you are running on the server like this?
And you can't predict/dictate a lot of things regarding the environment... which is why distributions for linux have everything + kitchen sink repositories. If wayland supports X11 apps on top of it (like OSX), and the gui toolkits (GTK, Qt) are migrated... I think that is a pretty good point of change. Most apps use GUI abstractions before the base X11 UI, so it should be easy enough to re-target them specifically.
I used to use a really nice LiteStep shell as a desktop replacement, liked it a lot.. it supported multiple desktops, with hot corners for swapping around. I stopped when I went to linux for a year, and then with win7, I liked it enough to stick with the default shell.
You can do Win8 UI apps in any.Net language as well. When they announced the ability to do desktop apps in JS + markup, there wasn't mention of.Net in the demo, people made assumptions and the FUD-fest was started. Personally, I like JS a lot, so despite liking C# as well, would probably go that route for anything new that had to do Win8. But then again, I've been doing a lot of Node.js development lately.
That's a given, and if the UI is heavily integrated with the business logic, it makes it far more difficult... I'm glad I work on web-based, and server applications (mostly).
WTF are you talking about... there's dozens of addons for windows to give you multiple workspaces.. even full on explorer shell replacements if you really don't like the UI.
So is supporting a codebase the size of an aircraft carrier, when nobody uses any of it beyond the flight deck... some times code bloat is so heavy, you need to make dramatic changes.
Because most window managers and gui toolkits in use today can't effectively use most of the spec, except bit block support and network streaming.. doing both together in X is often more overhead than an improved VNC could do... remote rendering is cool, but they aren't using the primitives in place. Why shouldn't we use an aircraft carrier instead of a speedboat? What 99.999% of what people need is supported by Wayland... The missing network support will likely be an application/wm abstraction to it, or added in later. RDP runs pretty damned well compared to VNC, I'm sure wayland can do something similar.
Slashdot Mirror
Explorer.exe is the entire desktop shell, not just the file browser.
Which is why I think there should be a pay floor of 125K/year on H1-B visa workers.. if there's nobody here to fill that job, it must require someone special, which means the pay should be that much higher or more.
Well, an experienced dev is likely to have 3-10x the output with a fraction of the bugs or other rework. Even the math works in the favor of a more experienced dev.
I usually spend the first couple hours 2-3 days a week reading blog posts, and keeping up... there are a huge amount of things to play with, choices seem to double every year, and many trends or toy frameworks don't make it. It's impossible to know everything in depth.. but it is entirely possible to spot trends, and learn what is becoming more common place. Right now, if you aren't at least looking at Node.js or Python, and contemplating a NoSQL database, you are going to be behind in a few years. Java and C#.Net are tried and true, and will be around for a while. Hell C/C++ and COBOL are still really broadly used. They're all needed and serve a purpose.
Ironically I think what was old is new again... That scripting environments and various data store options are coming to the front again. Scripting runtimes, and computing power are better than ever, and SQL doesn't fit every need well. I think in another 10 years there will be new dev environments... One size doesn't fit all... There is no perfect hammer... but what we can look for are better fits for a given need.
My best suggestion to a younger developer is don't get locked into one thing... think about the problems you are tasked to resolve and know your chosen environment as well as possible to resolve those problems. Keep your eyes open to new things... try a new language every year.. there's usually a couple to choose from. You don't have to do anything meaningful, or in the workplace to glean knowledge from something new. If something doesn't fit, don't force it.. and if something works better, push for it... don't get caught up in politics, and choose wisely.
I've seen it go both ways.. I'm pushing 40, and do a pretty good job keeping up, and being aware of new ideas, concepts and tools... on the flip side, I had a coworker that was a couple years older absolutely resistant to any new tools, techniques or change in general... Then again, I've seen plenty of younger developers who are relatively fresh from college who can't think their way through, under or around a given problem.
I've found that more and more a kanban approach tends to make people a little happier... tasks come in, tasks go out... the next task is prioritized, and interaction with others during a coding session should be mitigated by your manager... I'm also finding a unix-like approach of small bits of software that can pump in/out logic separately are a bit better than monolithic tightly coupled libraries... less OO, and more functional piped abstractions. Today I'm doing more of this in Node.js, it lets me get stuff done, usually with less code, tooling and overhead. You could do similar in C/C++, but that brings a compilation step into the mix. When I have to fire up an IDE and do a build to check/test/run/change anything, it just seems like overhead that shouldn't be needed anymore.
Using unix-like principles, and node (or any other scripted language) lets me get stuff done... following conventions make it more maintainable. I'm 37 and still learning... I can imagine I'll still be doing so for another few decades.
True enough.. I had both a Chevy Silverado 1500 crew cab, and a Hyundai Sonata before my current car a Challenger RT... I tended to do about 5-6mpg higher than rated in the sonata, and about the rating in the silverado... about 2/3 my driving was highway at the time... in my challenger, it's a bigger difference.. I tend to do best around 75mph relatively level than my current drive, about 1/2 highway, and heavy traffic... my general driving is around 19-20mpg, but I made a fairly decent trip a few weeks back averaging around 70-80mph, and got nearly 28mpg. How you drive has a big impact.. but traffic kills any predictability with fuel economy.
Honestly, the few times I've tried using google docs, the biggest hindrance imho is the lack of proper tabstop alignment.
I don't load most apps that require internet access, or for that matter contacts etc... there are few exceptions, but those are for apps that are related to the permissions they are asking for. I'm doubly leery when an app doesn't auto update, usually indicates a permission change.
given that my actual bank password cannot be anything other than us letters and numbers, no special characters (at two banking institutions)... An incident where I had over a million dollars in my account (magically) for a couple days, another where a bank error cost me a few hundred, and another still when a merger lost a friend over 10K with no tracking ability... I don't trust banks all that much either.
But consumption has risen since the effect of piracy (as long as you include online sales. (umn, err licenses, apologies eminem)
That doesn't make it a bad decision to stick with it... I find that more recent versions of gnome/unity/kde can be very sluggish on even modern hardware. I just want a usable desktop with a modest amount of eye candy that runs well. I really like the Win7 taskbar the best currently... Would say that the Mint desktop is probably a distant second, followed by osx. It's sad MS had to f*ck it all up with Win8.
As much as I don't even care for the surveillance aspects... It's highly unlikely that local law enforcement will have weaponized drones any time soon... In much the same way as they don't typically carry RPGs.
My first SSD had the bug.. was painful to eperience... been pretty happy with my corsair drives since then.
That isn't nearly so true... If you've used many ARM devices, there are a lot of projects that take a little, to a lot of tweaking for a non x86 environment.
Why have gui programs on your servers to begin with? I keep hearing this touted as the biggest reason for keeping X over Wayland, but still fail to see WTF you are running on the server like this?
And you can't predict/dictate a lot of things regarding the environment... which is why distributions for linux have everything + kitchen sink repositories. If wayland supports X11 apps on top of it (like OSX), and the gui toolkits (GTK, Qt) are migrated... I think that is a pretty good point of change. Most apps use GUI abstractions before the base X11 UI, so it should be easy enough to re-target them specifically.
I used to use a really nice LiteStep shell as a desktop replacement, liked it a lot.. it supported multiple desktops, with hot corners for swapping around. I stopped when I went to linux for a year, and then with win7, I liked it enough to stick with the default shell.
You can do Win8 UI apps in any .Net language as well. When they announced the ability to do desktop apps in JS + markup, there wasn't mention of .Net in the demo, people made assumptions and the FUD-fest was started. Personally, I like JS a lot, so despite liking C# as well, would probably go that route for anything new that had to do Win8. But then again, I've been doing a lot of Node.js development lately.
That's a given, and if the UI is heavily integrated with the business logic, it makes it far more difficult... I'm glad I work on web-based, and server applications (mostly).
Other than Flex based applications, I'm pretty sure it is a minority of web applications... (Note, I say applications, not sites).
I don't have any gui tools on my linux servers... so don't think I need the X11 compatibility.
WTF are you talking about... there's dozens of addons for windows to give you multiple workspaces.. even full on explorer shell replacements if you really don't like the UI.
So is supporting a codebase the size of an aircraft carrier, when nobody uses any of it beyond the flight deck... some times code bloat is so heavy, you need to make dramatic changes.
Because most window managers and gui toolkits in use today can't effectively use most of the spec, except bit block support and network streaming.. doing both together in X is often more overhead than an improved VNC could do... remote rendering is cool, but they aren't using the primitives in place. Why shouldn't we use an aircraft carrier instead of a speedboat? What 99.999% of what people need is supported by Wayland... The missing network support will likely be an application/wm abstraction to it, or added in later. RDP runs pretty damned well compared to VNC, I'm sure wayland can do something similar.