Forgot your password?

Comment: Re:The web is not a runtime environment. (Score 1) 586

by tlambert (#47434345) Attached to: Normal Humans Effectively Excluded From Developing Software

Basically, if you are thinking your browser is a "platform", or you are thinking "the web" is "a platform" in the traditional programming sense, as the OP obvious is, then you are an idiot.

No, actually, he's quite right. It's a different method of programming, a different paradigm altogether. He didn't talk about programming the browser so that part of your statement is irrelevant, but as a design platform the web truly is different. At least before people tried to change a markup language into a full page layout and presentation language.

The problem with web applications - and the intrinsic problem of abstraction of the complexity that's solved by historical runtime environments that the OP likes, is that the render is independent. The whole article the other day about the Google device lab:

Completely and totally underscores the fact that markup and rendering are separate from each other, and that the system doing the markup has to understand, and either have variant code that it outputs so that it renders the same in as many browsers as possible -- or you need an entire device lab, because you've given up on solving the problem, and are willing to employ someone other than a "Normal Human" (per the current article) in order to chip away on a per device basis, until you exhaustively cover all possibilities.

The separation on the render is the problem with the web, as a platform, and it's why it's * not* "a platform", it's "N back ends * M browsers" number of platforms.

This separation is the same mistake that was made when window management was separated from X windows, such that you didn't get the same look and feel on all applications based on having a particular X Terminal/X Server on which the render took place. In other words, the primitives were too primitive, and you ended up drawing boxes and lines and patterns, instead of "pop up menus" and "menu bards" and "dialog boxes".

What the OP in this article is bemoaning as being missing is a self-enforcing emergent property of the design decision to separate rendering from markup, and to separate markup from UI logic, and separate business logic from everything else. It's why web services are so complicated, and why they are so fragile.

The only thing that ever came close to dealing with the issue overall, at a high level, was WebObjects, and even then, it didn't try to do it in a way that was renderer/backend/middleware/security model/web server agnostic.

So again, I'm going to say that web services isn't a *platform* in the traditional sense of a computer running one of half a dozen 80x24 block mode terminals to front end a COBOL program was a platform, and that anyone who thinks it is ... is an idiot. At best, they are engaging in wishful thinking, if they think Microsoft, Oracle, IBM, and other vendors of these things are going to settle on a common programming paradigm, and turn themselves into commodities, which would result in about 1/6th the revenue they're getting today.

Comment: Re:Come now. (Score 1) 102

by tlambert (#47417945) Attached to: How Japan Lost Track of 640kg of Plutonium

Let's not make a big deal out of this. 640kg of reactor-grade plutonium is only enough for a bit over 100 fission bombs / fusion bomb first stages, merely enough to make the recipient roughly tied for being the world's sixth most armed nuclear power.

Nothing to see here.

Clearly, you have never built a fission device, if you think you could get that many of them out of 640kg of even weapons grade Plutonium. You need to probably go back and read "The Curve of Binding Energy" and recalculate the neutron numbers to determine critical mass, assuming a pareto optimal design, because you are more than a bit high with "100"...

You could build a lot of dirty bombs with something like that, but you are likely better off just robbing a radiomedicine unit at a large research hospital to get the materials, or stealing a truck out of a fast food restaurant in Mexico City...

Comment: The web is not a runtime environment. (Score 0) 586

by tlambert (#47417745) Attached to: Normal Humans Effectively Excluded From Developing Software

You are right of course it is similar to the 80's and 90's in that companies that wanted to steal the sales of other companies simply created new fangled languages and marketed the hell out of them instead of embracing what works and adapting it to the new paradigms. The only reason you can't use Turbo Pascal to make web pages is the compiler was never updated for the functionality but it very well could have been. In fact its progeny Delphi is alive and well and building apps for almost every popular platform out there today including the web. As long as there is competition there will be someone who chooses to create from scratch rather than use someone else's tool.

The web is not a runtime environment.

The reason you can't use TurboPascal is because web pages run in the browser virtual machine, and TurboPascal code runs in the TurboPascal runtime environment linked into the native code TurboPascal application.

You could target TurboPascal to NACL/PiNACL in Chrome as a target runtime environment, but effectively to run it, you'd be doing a JavaScript call into a JavaScript extension that then ran as native code in a sandbox within Chome. You'd, as a result, lose most of the TurboPascal runtime libraries supplied by the compiler vendor, and you'd lose all third party libraries and components, if the third parties weren't willing to port them (I assume you realize that you don't have all the Photoshop plugins on Windows that are available on Mac, right?).

Web languages, n the other hand, are predominantly for programming code on a server to generate markup, which is then interpreted by the browser to render output, or they are intended to run in a really limited environment in the browser itself, usually as unextended JavaScript (and, in the case of things like iPad/iPhone/etc., they are *definitely* NOT extended, since a UIView extension is not allowed under the terms and conditions for interpreting web content, since it's a huge security hole that's easily exploited with a DNS hijack).

Basically, if you are thinking your browser is a "platform", or you are thinking "the web" is "a platform" in the traditional programming sense, as the OP obvious is, then you are an idiot.

Comment: Better add DARPA and Jon Postel as codefendants (Score 2) 309

by tlambert (#47417297) Attached to: Tor Project Sued Over a Revenge Porn Business That Used Its Service

Better add DARPA and Jon Postel as codefendants. I hear they came up with this thing called TCP/IP, which aids and abets people like Tor putting together anonymous networks in the first place; it's a clear case of collusion...

Bonus Points: I hear DARPA has deep pockets...

Comment: Re:Virtual machines (Score 1) 60

by tlambert (#47410849) Attached to: All Web Developers Should Have Access to a Device Lab (Video)

What exactly is the point of spending so much money on hardware when you could run >40 virtual machines emulating different Android devices?

Most companies producing devices with browsers are pretty ass about providing working simulators/emulators for the hardware.

This is OK for one company, like Google or Roxio, to deal with supporting a lot of platforms with all sorts of physical differences from there being no hardware standard for Android devices to which vendors have to adhere, but ... it's not going to address the underlying problem, just because you can make the render device variant with less effort.

Comment: FWIW, Washtech is a CWA union local... (Score 1) 401

by tlambert (#47397619) Attached to: No Shortage In Tech Workers, Advocacy Groups Say

FWIW, Washtech is a CWA union local...

It's possible that they have the best interests of IT people in their hearts, but it's more likely that they, like the Alliance@IBM guys, also a CWA union local, have a bit of an axe to grind against IBM.

The other two seem more or less non-affiliated, so they perhaps do not have an axe to grind against IBM. It'd bee interesting to know which group(s) picked which target(s) in this story.

Also, FWIW, the CWA is a pretty piss poor match for programmers and other IT folks, but since automation of telephone operators jobs, they've been branching out to "anyone who uses a communications network, no matter how automated and non-labor intensive" as potential members. It's not a great fit, so they've had pretty much zero success in the IBM shops they've picketed (including one I worked for at one time).

Comment: Re:Two sides to every issue (Score 1) 401

by tlambert (#47397613) Attached to: No Shortage In Tech Workers, Advocacy Groups Say

H1-Bs in America currently have two options: 1) Remain at current sponsoring employer or 2) go home, because quitting means immediate revocation of their visa.

2B: Hop to an employer that is willing to sponsor a change in their H1-B.

From Wikipedia:

Despite a limit on length of stay, no requirement exists that the individual remain for any period in the job the visa was originally issued for. This is known as H-1B portability or transfer, provided the new employer sponsors another H-1B visa

From the employees perspective, there is one problem with this: once an employer has started the permanent residency (greencard) process, it is a bad idea to move because you'll be starting all over again.

A take-over is easier than a reapplication for a new visa, if the current visa limit is exhausted (which it constantly is), so unless this happens at the start of a year, and you have all the ducks in a row before tendering notice, you are likely going home as soon as you give notice to the current visa sponsor.

A take-over is allowed, but voluntary on the part of the original sponsor, who may be, er, a "little spiteful"...

The end of labor is to gain leisure.