Warning: multiple transition points
Late at night in the messenger, a much younger programmer asked me: “Damn, is there so much happening, is it always like that?” Hmm, no, it wasn’t like that before. But so may be in the future.
As far as I can tell, we are now seeing several transition points at the same time in programming languages, and databases, and network programming, and processor architectures, and web development, and IT business models, and work environments. Did I miss anything? But even more interesting is that such a regime of multiple transition points may continue for several more years.
')
Until very recently, you would choose Java, or .NET, or, if you're a real masochist, C ++ for a big serious software project. And today you will be a fool if you do not look closely at PHP, Python and Ruby. How much attention will they attract in the future? I do not know, but this decision is made by the whole community right now .
By the way, on this topic, look at the Interview with the TIOBE guy - an interesting thing, although I do not agree with everything in it.
Couchdb . Simpledb . BigTable . Is there anything else to say?
No, I do not think that relational databases will disappear soon. But I think that the closure of the SQL developer community has hurt them a lot over the past couple of decades, and I am glad that it is now considered normal to look at alternatives.
Will non-relational alternatives tear off a piece of the market for themselves? I think so, but this decision is made by the whole community right now .
CORBA is dead. DCOM is dead. WS- * already hacking cough on the way to the grave. They say REST is the right way. And I agree with that. But he still has few tools, and best practices, and accumulated wisdom, and consultants in blue suits, and other signs of mainstream technology.
So how in a few years will children be taught the right way to create an application on the net, full of heterogeneous technologies? This is decided by the whole community right now .
Moore's law still works, but processors grow in breadth, rather than becoming faster. After the best of the best have spent a decade on creating and debugging threading frameworks in Java and .NET, it becomes increasingly clear that this is a bad idea, enough to develop it. I myself completely changed my previous pro-streaming position after I came to Sun four years ago.
We have not yet figured out the correct programming method for multi-core processors for an ordinary person — check the inconclusive results of my last year’s project Wide Finder . (By the way, I now have my own T2000 exposed on the Internet, and as soon as there is enough data on it, I will restart Wide Finder; come all).
I can’t even repeat here my suggestion about the right answer, which is being sought right now , because I’m not really sure that at least someone understands what is needed here. But we are sure that we are passing the transition point here.
Previously, it was Java EE, or Perl, or ASP.NET. And now, unexpectedly, this is PHP, and then Rails and a whole bunch of frameworks rising on the horizon; and the month does not pass, so that I do not see another hype about something "similar to Rails".
It seems to me obvious that Rails ++ will appear very soon, which will combine good ideas from RoR and some others that are obvious right after you see them for the first time.
And also that some of these “rail-like” frameworks, even if they do not make any breakthroughs, will still find their market share, because they combine several minor advantages.
Again, I can't say that it is becoming clear right now , because right now I see an unequivocal picture of the steadily growing Rails market. But it will not be long.
On the servers, everything is simple and clear. Blue-suited salespeople sell servers to technical directors in portions of hundreds of thousands of dollars, they are loaded into data centers, where they require too much electricity and maintenance.
Rather, it is if you want to mess around with storage, computing power and load distribution yourself, and not send it all to the cloud. Want to? Technical directors and data center guys are dealing with this issue right now .
As for the software, before it was sold on magnetic media, and the money was taken for the right to use. Now everything is in open source, and they are downloaded for free, and you provide paid technical support. Although no, it was also in the last century, perhaps now all the software will be somewhere in the cloud, and you will not even download anything, just pay for the use.
I personally do not think that any of these models will completely disappear. But which one works best? The market decides this right now .
As I wrote a few months ago: how much time will public and private technical management continue to ignore the fact that there is OS X and Ubuntu, that there are even two alternatives to the Windows environment that are safer, safer, more efficient and cheaper? Probably, now everyone has seen a poppy or Linux from a friend or relative, and thinks why his computer cannot be just as cool.
What will happen here? I do not know, but it will be a dramatic change when we find ourselves at the transition point, and we are approaching it right now .
You know, maybe. Our ownership of information technology is still very young, and many fruits hang very low, and many broad steps forward are waiting for them to be made. Now, with the proliferation of blogs, and non-conferences, and all these new ways of communicating , our thinkers constantly, 24/7/365, communicate with each other about all these problems. The gap between the leading edge and the technology really used in corporations is still as huge as before, and it seems to me that it is thanks to him that continuous destruction occurs. Cobangang!
Warning: multiple transition points @ blog.arty.name
As far as I can tell, we are now seeing several transition points at the same time in programming languages, and databases, and network programming, and processor architectures, and web development, and IT business models, and work environments. Did I miss anything? But even more interesting is that such a regime of multiple transition points may continue for several more years.
')
Programming languages
Until very recently, you would choose Java, or .NET, or, if you're a real masochist, C ++ for a big serious software project. And today you will be a fool if you do not look closely at PHP, Python and Ruby. How much attention will they attract in the future? I do not know, but this decision is made by the whole community right now .
By the way, on this topic, look at the Interview with the TIOBE guy - an interesting thing, although I do not agree with everything in it.
Database
Couchdb . Simpledb . BigTable . Is there anything else to say?
No, I do not think that relational databases will disappear soon. But I think that the closure of the SQL developer community has hurt them a lot over the past couple of decades, and I am glad that it is now considered normal to look at alternatives.
Will non-relational alternatives tear off a piece of the market for themselves? I think so, but this decision is made by the whole community right now .
Network programming
CORBA is dead. DCOM is dead. WS- * already hacking cough on the way to the grave. They say REST is the right way. And I agree with that. But he still has few tools, and best practices, and accumulated wisdom, and consultants in blue suits, and other signs of mainstream technology.
So how in a few years will children be taught the right way to create an application on the net, full of heterogeneous technologies? This is decided by the whole community right now .
Processors
Moore's law still works, but processors grow in breadth, rather than becoming faster. After the best of the best have spent a decade on creating and debugging threading frameworks in Java and .NET, it becomes increasingly clear that this is a bad idea, enough to develop it. I myself completely changed my previous pro-streaming position after I came to Sun four years ago.
We have not yet figured out the correct programming method for multi-core processors for an ordinary person — check the inconclusive results of my last year’s project Wide Finder . (By the way, I now have my own T2000 exposed on the Internet, and as soon as there is enough data on it, I will restart Wide Finder; come all).
I can’t even repeat here my suggestion about the right answer, which is being sought right now , because I’m not really sure that at least someone understands what is needed here. But we are sure that we are passing the transition point here.
Web development
Previously, it was Java EE, or Perl, or ASP.NET. And now, unexpectedly, this is PHP, and then Rails and a whole bunch of frameworks rising on the horizon; and the month does not pass, so that I do not see another hype about something "similar to Rails".
It seems to me obvious that Rails ++ will appear very soon, which will combine good ideas from RoR and some others that are obvious right after you see them for the first time.
And also that some of these “rail-like” frameworks, even if they do not make any breakthroughs, will still find their market share, because they combine several minor advantages.
Again, I can't say that it is becoming clear right now , because right now I see an unequivocal picture of the steadily growing Rails market. But it will not be long.
Business models
On the servers, everything is simple and clear. Blue-suited salespeople sell servers to technical directors in portions of hundreds of thousands of dollars, they are loaded into data centers, where they require too much electricity and maintenance.
Rather, it is if you want to mess around with storage, computing power and load distribution yourself, and not send it all to the cloud. Want to? Technical directors and data center guys are dealing with this issue right now .
As for the software, before it was sold on magnetic media, and the money was taken for the right to use. Now everything is in open source, and they are downloaded for free, and you provide paid technical support. Although no, it was also in the last century, perhaps now all the software will be somewhere in the cloud, and you will not even download anything, just pay for the use.
I personally do not think that any of these models will completely disappear. But which one works best? The market decides this right now .
Working environments
As I wrote a few months ago: how much time will public and private technical management continue to ignore the fact that there is OS X and Ubuntu, that there are even two alternatives to the Windows environment that are safer, safer, more efficient and cheaper? Probably, now everyone has seen a poppy or Linux from a friend or relative, and thinks why his computer cannot be just as cool.
What will happen here? I do not know, but it will be a dramatic change when we find ourselves at the transition point, and we are approaching it right now .
Will it always be like this?
You know, maybe. Our ownership of information technology is still very young, and many fruits hang very low, and many broad steps forward are waiting for them to be made. Now, with the proliferation of blogs, and non-conferences, and all these new ways of communicating , our thinkers constantly, 24/7/365, communicate with each other about all these problems. The gap between the leading edge and the technology really used in corporations is still as huge as before, and it seems to me that it is thanks to him that continuous destruction occurs. Cobangang!
Warning: multiple transition points @ blog.arty.name
Source: https://habr.com/ru/post/31529/
All Articles