At it's best it's negligibly faster, at it's worst it's negligibly slower. They are in the same ballpark, playing the same game, the only thing is that Python makes playing the game more fun and better experience. If you ever need that raw data crunching performance Python today has much better ecosystem of wrapped C libraries, Cython, PyPy etc. Maybe CPAN is huge but most people won't bother as everything around it is kinda arcane. And then we come to the point that, as extension of all this, many more people are adding performance tweaks to NumPy than to Math::Pari or Math::GSL.
I mean, numbers (GitHub repos, developer users, hell even benchmarks) speak for itself so you can pretend it's not happening but it is.
Perl has, well, to say gently, less than easily manageable syntax, that was still saner than Awk and Sed. So when it became ubiquitous on most systems many complex scripting has migrated from sh/bash+awk+sed to a much saner, more developer-friendly Perl interpreter, which then started to be used for scripting webpages as that was (at the time) the next big thing.
Then came PHP and most new users never learned Perl to work solely with websites (and most never really got down-dirty to *nixen either as WAMP thingies enabled all that jazz on Windows).
Then Python became ubiquitous on most systems, and because it provides an even nicer and cleaner development experience people generally started using Python for everything they used Perl for, then more and more often for things they used PHP for, and then Data Science thing came and Python ecosystem was very quick and efficient in embracing it so it started to become the dominating language of choice for quick-fixes and one-offs and as extension inevitably for more and more large, business-critical software development, as tt's often the case thus becoming an underdog competitor to Java much like PHP was/is.
Now there are tastes and tastes but there is still some, if not generality, then at least majority or dominance. Sure, some people really love VB, but most don't. Sure, some people really hate Python but most don't.
Oh deary me. In fact for $1000 XPS13 gets you a 5200U scoring 3800 at cpumark, and at $1300 Mac gets you twice as big SSD and a 5257U scoring 4250. You can probably invest those 300 in lots more than more SSD with Dell. And you get a leaner machine. Also, Windows and Ubuntu will fly on the Dell. OSX is a horrible hog (that's why Macbooks always have beefy processors, even though most of their users only use them for facebook, iworks and email), and everyone using Windows or LInux on them will testify that they simply fly, whereas people running Hackintoshes on mid-spec notebooks aren't having nearly as much fun despite their machines being very performant with WIndows, even with all that bloatware users typically get, and especially Linux.
PyPy.. Obviously not as fast as HotSpot or V8 but in the league at leas. OTOH using it still means you have to give up on many binary extensions due to JIT nature of PyPy which requires different bindings.
Java is.. not very good. I've alternated between Java and C# professionally over the past 8 years or so, and while they used to be quite similar, C# is worlds ahead now (thanks, Oracle!).
I'm no fan of Java, but I'd really like to hear about where C# is worlds ahead of Java now. We've had "evangelist gone sales rep" from local Microsoft in our Java/OracleDB shop recently. I was somehow in-between.Net open sourcing and first Mono based on.Net Core release, so we were honestly interested, esp. since purchasing VS licences was an option, compared to running everything on WinServer, which simply wasn't.
What was shown to us was visibly catching up, but still behind modern MVC frameworks, support for version control systems was also quite behind (I can assume TeamServe or what's it called must look like sci-fi to people who never used modern project management and CI software, but it's not) what is currently available in all FLOSS dev ecosystems, including Java (which, granted, is the most conservative one). Both are behind the curve compared to RoR, Python and node.js, and MVC6 is trying hard to be there where modern Node frameworks are (MS is very openly interested in node)... All in all, we weren't impressed very much.
I can see few more plausible reasons for the decline:
1) Servers are increasingly Linux, even in enterprise. Microsoft response by open-sourcing.Net might be too little, loo late. 2) Client apps are increasingly mobile, Microsoft's market-share here is dismal.
1) and 2) have given boost to Java, boost to Java is instantly C# decline since both languages rely on same approach to project management (the "nobody ever got fired for...." style) to thrive.
3) Visual Studio my be the bees knees, but.Net is far, far behind the curve when it comes to web application development, which brings us to: 4) Start-up economy is on fire, and there it's a run for least verbose, fastest to market type of development (release early, release often, pivot often, get tons of cash, then rewrite backend in something that can actually scale --- see: Twitter). Success in software-development related economy is not (and frankly, never was) about most solid technology choices from the get go, and more and more devterpreneurs are seeing this.
Re 3) and 4) I can't think of any major web app (apart from StackExchange, granted) that is based on a Microsoft stack.
There is no such thing as file type flag in OSX tho, it uses the NeXT concept of bundle directories (which NeXT actually borrowed from RiscOS). The extension that marks applications is.app and "app bundle" is quite common name for it. Pretty sure it's where the "app" shortening for "application" got it's ubiquitousness that then spread with popularity of iOS.
But I distinctively remember even back in the 90s that the term Application was the "proper" term for software installation - as Application can consist of multiple Programs, but not vice versa.
Well no, it is not. Productivity cannot be measured in GDP due to various (political, military, ownership/control of resources etc.) factors skewing international trade that it's derived from.
Anyway, Americans are obviously not top country vis-a-vis GDP. And that is after the fact that USA is the big-dog of international relations that doesn't shy from imposing control over someone else's resources using military aggression (Afghanistan, Iraq, Libya to name a few), which is hardly true for places like Luxembourg or Norway.
I was not talking about state-owned telcos, only state-owned body that leases the core infrastructure and leases it to telcos on equal basis. Because telcos often own the core infrastructure, it gives those big (and often previously state-owned telcos) unfair advantage over smaller, also private telcos, in addition to other concerns stated above. When these are privatized or stock marketed former-state telcos it also means that someone (not the public) is getting the infrastructure that was funded from taxpayers money.
Much like there are no private owned roads, highways or railways, there should be no private owned backbones, and telcos should find a way of private/public parnership programmes if they want to incentivize technology-wise progress of the backbones, or utilize what is already there.
Except that TFA clearly states they are porting Ozone (their graphics toolkit, as they're not using Qt Gtk or anything else) to Wayland.
Common missunderstanding is that Wayland is a display server. It is not. It's a protocol that UI toolkits use to talk to the compositor (KWin, Weston etc.).
I'm pretty sure that Canonical, by now, are more or less certian they will need to provide for apps to talk Wayland to Mir as not everyone will be using one of the big UI toolkits that talk to Mir directly. To those apps, Mir will be just another compositor.
And I don't mean Google, Google is pragmatic enough and has manpower to port Ozone to Mir as well.
3: as in most countries in the world the core telecommunication infrastructure was government, tax-sponsored investment, which was then put on the stock market together with telecom providers, they could take back that infrastructure (which should have been leased to telcos in the first place), and control it through regulatory bodies for telecom, and lease that to telcos. Then they could technically enforce routing rules (and frankly, I still fail to see how it's better for freedom if corporations control our wires vs. when government does).
"Furthermore, with the continued erosion of the US economy into a non-producing one,"
Really? The latest widely-published statistics I read said that the average American is still 30% more productive per work hour than the rest of the Western world.
I know it's unpopular to say that, but if there weren't global-level pressures from socialist organizations you'd get fsckall of those 40-hour weeks and work safety. Unions solved (and still do) issues on trade by trade basis. Overall conditions of workers improved only when powers that were felt grass roots pressure from protesting and increasing number of people going the red route everywhere. The whole red scare thing was more-less designed to create a stigma over a whole concept of labour rights in the West, leaving trade Unions to become charades quite often.
Tho, charming personalities like Stalin and Mao helped a lot. Nothing says an idea is broken better than pointing at a perverted, evil implementation of it.
It's always interesting to read how the real heroes, those that made real contributions, were warm, pleasant personalities. Many people described Dennis Ritchie with similar words. It's such a stark contrast to unpleasant, egotistical, NPD personalities that make most of the headlines.
RIP
What vendor support contract will get you your personal bug fix-it team? You obviously never worked with Microsoft and their support, or you're in a Fortune 500 that can spend couple of millions yearly to get that level of support (i.e. you're one of the reasons why regular users get regression bugs or idiotic functionality since you have the money to make the vendor cut updates to suit your fancy).
In the real world (of small and medium companies, you know, where most IT people actually work), your best served by a small vendor, preferably local, that considers YOU an important customer. And provided that he employs some developers himself for him the best model is Open Source since he can tailor it to his customers needs.
Perhaps... but I beleive it's the concentration of open-walleted investors and the whole VC concept that made Sillicon Valley what it is. There are smart people everywhere in the world. If concentration of uber-smart people was THE requirement it would have been, say, Mumbai, but Indians are skint and famously mingy so it's not. Those brilliant global PhDs wouldn't concentrate in SoCa if weren't for the money that was very willing to mobilize.
It's still so. It's still much easier to get funding if your start-up's addy is in the vincinity of SF bay area even tho you can do great things everywhere. Sadly, in many cases underfunded ventures don't go anywhere. There is a whole history technically superior products originating outside SV and failing mostly due to underfunding. The times are changing tho, internetization and production outsourcing to China democratized the market, and OTOH many policy makers in EU for example are understanding the need to angel fund tech startups, which is why there are now more successful tech companies coming from EU than ever.
My office has a Macbook Pro, three Windows boxen (two on XP, one on 7), two Linux Mint boxes (a desktop and a laptop) and three Xubuntu desktops. Two CentOS servers running VMs for our business software, a Linux based VOIP PBX, a Linux based router, a Linux based WiFi box.. one iPhone, four Android phones, my old defunct HTC Beatle (iPaq 6515) was running WinCE while it was working tho. All windows boxen run Libre 3.5, we're cheap like that.
But here is pirated-Windows land. Just a couple hundred miles from here is EU. Anecdotaly, a friend of mine works for a small Linux support company in Germany. He says that the number of Linux desktops (and office servers but that was the trend even before) has dramatically increased post 2009 recession.
And yes, Apple probably sells more laptops than any single laptop brand in the world right now.
eMagic was also bought when Logic Audio was a pro/prosumer studio standard, in some parts of the world even more so than Steinberg's or Avid's (ProTools) software. Then they released Garage Band partly based on the same engine but with a more consumer-oriented UI. Then they slowly started adjusting the UI and the very model of Logic (does anyone remember those layouts you could design in Logic) to fit consumer mindset.
But it was for the better in terms of sales, and hurt only a few uber high end users: Logic now competes in prosumer/consumer area with massively selling consumer/amateur software like FL Studio (typically disregarded by professionals but nowadays probably the best selling sequencer and used by large number of dance industries royalty, it earned the company owner a private jet among other things) and not only does it shift units, it also happens to sell quite a few Macintosh boxes to those interested as well.
So someone at Apple noticed the potential of audio prosumer/bedroom musician market and put some energy behind it, dumbed down an industry standard in order to expand the market and still capitalize on the brand, and I'm sure it adds a notch on Mac sales. They might not be no.1 in terms of sales but they are in terms of mindshare (a lot of Apple fanboys who consider only Logic is a proper tool everywhere).
I thought the original intention of patents was really to protect and enhance manufacturing.
The original intent of parents, i.e. not the one patent trolls and their lawyers advertise, was to protect knowledge from being lost/forgotten due to industrial secrecy. The intent was to convert private knowledge into public good in exchange for a limited-time legal protection. The idea was never really to ensure monopoly and especially not income of the inventor, but to prevent inventor from keeping his inventions secret by allowing him to be legally protected for a limited time in exchange for publicly disclosing his inventions.
Which is why software patents are retarded and esp. Apple-style trivial UI patents because the society as a whole doesn't benefit from obvious UI gimmicks being presented in a patent format since, even if they weren't non-trivial and derivative, they would be in plain sight and not a part of otherwise hidden internal functioning. So even if the reasoning behind math and business processes not being patentable are not analogous enough there is additional argument against sw patents.
So the entire system that enforces patents of the latter kind is either dishonest or ignorant about the purpose of patent system. I don't know which of the two is worse as a trait for a judge. At least Common law judges can hide behind precedents (but hide only, they are not as bound by precedents as they are by the intent and purpose of a legal institute), but in civil law jurisdictions it would be unforgivable. Which is methinks why Apple sued Samsung on industrial design rather than patents in Germany.
Slashdot Mirror
Whereas Perl is somehow performant?
At it's best it's negligibly faster, at it's worst it's negligibly slower. They are in the same ballpark, playing the same game, the only thing is that Python makes playing the game more fun and better experience. If you ever need that raw data crunching performance Python today has much better ecosystem of wrapped C libraries, Cython, PyPy etc. Maybe CPAN is huge but most people won't bother as everything around it is kinda arcane. And then we come to the point that, as extension of all this, many more people are adding performance tweaks to NumPy than to Math::Pari or Math::GSL.
I mean, numbers (GitHub repos, developer users, hell even benchmarks) speak for itself so you can pretend it's not happening but it is.
I think the whole thing boils down to utility.
Perl has, well, to say gently, less than easily manageable syntax, that was still saner than Awk and Sed. So when it became ubiquitous on most systems many complex scripting has migrated from sh/bash+awk+sed to a much saner, more developer-friendly Perl interpreter, which then started to be used for scripting webpages as that was (at the time) the next big thing.
Then came PHP and most new users never learned Perl to work solely with websites (and most never really got down-dirty to *nixen either as WAMP thingies enabled all that jazz on Windows).
Then Python became ubiquitous on most systems, and because it provides an even nicer and cleaner development experience people generally started using Python for everything they used Perl for, then more and more often for things they used PHP for, and then Data Science thing came and Python ecosystem was very quick and efficient in embracing it so it started to become the dominating language of choice for quick-fixes and one-offs and as extension inevitably for more and more large, business-critical software development, as tt's often the case thus becoming an underdog competitor to Java much like PHP was/is.
Now there are tastes and tastes but there is still some, if not generality, then at least majority or dominance. Sure, some people really love VB, but most don't. Sure, some people really hate Python but most don't.
Lua
Just did. Copy the folder (including the node_modules tree). Run `node .` (the dot is not a typo). Done
Oh deary me. In fact for $1000 XPS13 gets you a 5200U scoring 3800 at cpumark, and at $1300 Mac gets you twice as big SSD and a 5257U scoring 4250. You can probably invest those 300 in lots more than more SSD with Dell. And you get a leaner machine. Also, Windows and Ubuntu will fly on the Dell. OSX is a horrible hog (that's why Macbooks always have beefy processors, even though most of their users only use them for facebook, iworks and email), and everyone using Windows or LInux on them will testify that they simply fly, whereas people running Hackintoshes on mid-spec notebooks aren't having nearly as much fun despite their machines being very performant with WIndows, even with all that bloatware users typically get, and especially Linux.
PyPy.. Obviously not as fast as HotSpot or V8 but in the league at leas. OTOH using it still means you have to give up on many binary extensions due to JIT nature of PyPy which requires different bindings.
And you weren't even a customer!
Java is .. not very good. I've alternated between Java and C# professionally over the past 8 years or so, and while they used to be quite similar, C# is worlds ahead now (thanks, Oracle!).
I'm no fan of Java, but I'd really like to hear about where C# is worlds ahead of Java now. We've had "evangelist gone sales rep" from local Microsoft in our Java/OracleDB shop recently. I was somehow in-between .Net open sourcing and first Mono based on .Net Core release, so we were honestly interested, esp. since purchasing VS licences was an option, compared to running everything on WinServer, which simply wasn't.
What was shown to us was visibly catching up, but still behind modern MVC frameworks, support for version control systems was also quite behind (I can assume TeamServe or what's it called must look like sci-fi to people who never used modern project management and CI software, but it's not) what is currently available in all FLOSS dev ecosystems, including Java (which, granted, is the most conservative one). Both are behind the curve compared to RoR, Python and node.js, and MVC6 is trying hard to be there where modern Node frameworks are (MS is very openly interested in node)... All in all, we weren't impressed very much.
I can see few more plausible reasons for the decline:
1) Servers are increasingly Linux, even in enterprise. Microsoft response by open-sourcing .Net might be too little, loo late.
2) Client apps are increasingly mobile, Microsoft's market-share here is dismal.
1) and 2) have given boost to Java, boost to Java is instantly C# decline since both languages rely on same approach to project management (the "nobody ever got fired for...." style) to thrive.
3) Visual Studio my be the bees knees, but .Net is far, far behind the curve when it comes to web application development, which brings us to:
4) Start-up economy is on fire, and there it's a run for least verbose, fastest to market type of development (release early, release often, pivot often, get tons of cash, then rewrite backend in something that can actually scale --- see: Twitter). Success in software-development related economy is not (and frankly, never was) about most solid technology choices from the get go, and more and more devterpreneurs are seeing this.
Re 3) and 4) I can't think of any major web app (apart from StackExchange, granted) that is based on a Microsoft stack.
There is no such thing as file type flag in OSX tho, it uses the NeXT concept of bundle directories (which NeXT actually borrowed from RiscOS). The extension that marks applications is .app and "app bundle" is quite common name for it. Pretty sure it's where the "app" shortening for "application" got it's ubiquitousness that then spread with popularity of iOS.
But I distinctively remember even back in the 90s that the term Application was the "proper" term for software installation - as Application can consist of multiple Programs, but not vice versa.
Well no, it is not. Productivity cannot be measured in GDP due to various (political, military, ownership/control of resources etc.) factors skewing international trade that it's derived from.
Anyway, Americans are obviously not top country vis-a-vis GDP. And that is after the fact that USA is the big-dog of international relations that doesn't shy from imposing control over someone else's resources using military aggression (Afghanistan, Iraq, Libya to name a few), which is hardly true for places like Luxembourg or Norway.
I was not talking about state-owned telcos, only state-owned body that leases the core infrastructure and leases it to telcos on equal basis. Because telcos often own the core infrastructure, it gives those big (and often previously state-owned telcos) unfair advantage over smaller, also private telcos, in addition to other concerns stated above. When these are privatized or stock marketed former-state telcos it also means that someone (not the public) is getting the infrastructure that was funded from taxpayers money.
Much like there are no private owned roads, highways or railways, there should be no private owned backbones, and telcos should find a way of private/public parnership programmes if they want to incentivize technology-wise progress of the backbones, or utilize what is already there.
No, Aura/Ozone is the toolkit (Aura being the widget/chrome layer, and Ozone lower level graphic layer).
Except that TFA clearly states they are porting Ozone (their graphics toolkit, as they're not using Qt Gtk or anything else) to Wayland.
Common missunderstanding is that Wayland is a display server. It is not. It's a protocol that UI toolkits use to talk to the compositor (KWin, Weston etc.).
I'm pretty sure that Canonical, by now, are more or less certian they will need to provide for apps to talk Wayland to Mir as not everyone will be using one of the big UI toolkits that talk to Mir directly. To those apps, Mir will be just another compositor.
And I don't mean Google, Google is pragmatic enough and has manpower to port Ozone to Mir as well.
3: as in most countries in the world the core telecommunication infrastructure was government, tax-sponsored investment, which was then put on the stock market together with telecom providers, they could take back that infrastructure (which should have been leased to telcos in the first place), and control it through regulatory bodies for telecom, and lease that to telcos. Then they could technically enforce routing rules (and frankly, I still fail to see how it's better for freedom if corporations control our wires vs. when government does).
"Furthermore, with the continued erosion of the US economy into a non-producing one,"
Really? The latest widely-published statistics I read said that the average American is still 30% more productive per work hour than the rest of the Western world.
[citation needed]
I know it's unpopular to say that, but if there weren't global-level pressures from socialist organizations you'd get fsckall of those 40-hour weeks and work safety. Unions solved (and still do) issues on trade by trade basis. Overall conditions of workers improved only when powers that were felt grass roots pressure from protesting and increasing number of people going the red route everywhere. The whole red scare thing was more-less designed to create a stigma over a whole concept of labour rights in the West, leaving trade Unions to become charades quite often. Tho, charming personalities like Stalin and Mao helped a lot. Nothing says an idea is broken better than pointing at a perverted, evil implementation of it.
It's always interesting to read how the real heroes, those that made real contributions, were warm, pleasant personalities. Many people described Dennis Ritchie with similar words. It's such a stark contrast to unpleasant, egotistical, NPD personalities that make most of the headlines. RIP
What vendor support contract will get you your personal bug fix-it team? You obviously never worked with Microsoft and their support, or you're in a Fortune 500 that can spend couple of millions yearly to get that level of support (i.e. you're one of the reasons why regular users get regression bugs or idiotic functionality since you have the money to make the vendor cut updates to suit your fancy). In the real world (of small and medium companies, you know, where most IT people actually work), your best served by a small vendor, preferably local, that considers YOU an important customer. And provided that he employs some developers himself for him the best model is Open Source since he can tailor it to his customers needs.
Perhaps... but I beleive it's the concentration of open-walleted investors and the whole VC concept that made Sillicon Valley what it is. There are smart people everywhere in the world. If concentration of uber-smart people was THE requirement it would have been, say, Mumbai, but Indians are skint and famously mingy so it's not. Those brilliant global PhDs wouldn't concentrate in SoCa if weren't for the money that was very willing to mobilize.
It's still so. It's still much easier to get funding if your start-up's addy is in the vincinity of SF bay area even tho you can do great things everywhere. Sadly, in many cases underfunded ventures don't go anywhere. There is a whole history technically superior products originating outside SV and failing mostly due to underfunding. The times are changing tho, internetization and production outsourcing to China democratized the market, and OTOH many policy makers in EU for example are understanding the need to angel fund tech startups, which is why there are now more successful tech companies coming from EU than ever.
My office has a Macbook Pro, three Windows boxen (two on XP, one on 7), two Linux Mint boxes (a desktop and a laptop) and three Xubuntu desktops. Two CentOS servers running VMs for our business software, a Linux based VOIP PBX, a Linux based router, a Linux based WiFi box.. one iPhone, four Android phones, my old defunct HTC Beatle (iPaq 6515) was running WinCE while it was working tho. All windows boxen run Libre 3.5, we're cheap like that.
But here is pirated-Windows land. Just a couple hundred miles from here is EU. Anecdotaly, a friend of mine works for a small Linux support company in Germany. He says that the number of Linux desktops (and office servers but that was the trend even before) has dramatically increased post 2009 recession.
And yes, Apple probably sells more laptops than any single laptop brand in the world right now.
Similar story in pro-audio.
eMagic was also bought when Logic Audio was a pro/prosumer studio standard, in some parts of the world even more so than Steinberg's or Avid's (ProTools) software. Then they released Garage Band partly based on the same engine but with a more consumer-oriented UI. Then they slowly started adjusting the UI and the very model of Logic (does anyone remember those layouts you could design in Logic) to fit consumer mindset.
But it was for the better in terms of sales, and hurt only a few uber high end users: Logic now competes in prosumer/consumer area with massively selling consumer/amateur software like FL Studio (typically disregarded by professionals but nowadays probably the best selling sequencer and used by large number of dance industries royalty, it earned the company owner a private jet among other things) and not only does it shift units, it also happens to sell quite a few Macintosh boxes to those interested as well.
So someone at Apple noticed the potential of audio prosumer/bedroom musician market and put some energy behind it, dumbed down an industry standard in order to expand the market and still capitalize on the brand, and I'm sure it adds a notch on Mac sales. They might not be no.1 in terms of sales but they are in terms of mindshare (a lot of Apple fanboys who consider only Logic is a proper tool everywhere).
How then do you explain the track record of one judge Lucy Koh in the Apple vs. everybody saga?
I thought the original intention of patents was really to protect and enhance manufacturing.
The original intent of parents, i.e. not the one patent trolls and their lawyers advertise, was to protect knowledge from being lost/forgotten due to industrial secrecy. The intent was to convert private knowledge into public good in exchange for a limited-time legal protection. The idea was never really to ensure monopoly and especially not income of the inventor, but to prevent inventor from keeping his inventions secret by allowing him to be legally protected for a limited time in exchange for publicly disclosing his inventions.
Which is why software patents are retarded and esp. Apple-style trivial UI patents because the society as a whole doesn't benefit from obvious UI gimmicks being presented in a patent format since, even if they weren't non-trivial and derivative, they would be in plain sight and not a part of otherwise hidden internal functioning. So even if the reasoning behind math and business processes not being patentable are not analogous enough there is additional argument against sw patents.
So the entire system that enforces patents of the latter kind is either dishonest or ignorant about the purpose of patent system. I don't know which of the two is worse as a trait for a judge. At least Common law judges can hide behind precedents (but hide only, they are not as bound by precedents as they are by the intent and purpose of a legal institute), but in civil law jurisdictions it would be unforgivable. Which is methinks why Apple sued Samsung on industrial design rather than patents in Germany.