Page 7 of 8 « First 56 7 8
Reply
Thread Tools
Radu is offline Radu
2011-10-03 , 23:53
Posts: 303 | Thanked: 146 times | Joined on Aug 2009
#61
Originally Posted by Joorin View Post
No, it takes quite a bit more to do the same thing. You need a canvas, you need font support, you need to tinker with window handling, you need a compiler and an environment for it to run in and you need to assemble all these parts to get a binary that writes "Hello world!" in a window.
Those are tools that all the C programmers that do GUI stuff already have.

That's my point. And one benefit with browser bound development is that you can do all this with a simple text editor and the "Reload" button to review the result.
Or you can do a compile and run, which is one button too (depending on your IDE). And obviously, if you use native code in your HTML5 app, doing a refresh won't do much..

I'm sure this is some ways away from the goal of HTML5.
Exactly. HTML5 has a different target and purpose (web mainly).

But as a suggestion, I'd venture to guess that streaming video will use the same native support libraries no matter if you use GTK or HTML5 for the decoding. Then you need to pay for parsing the HTML document once and then it's pretty much the same CPU cost.
Yes, streaming video is going to use native code, and it is one of the reasons why HTML5 was designed for. So that's a good use for it.

You wrote the following:
But why? How many people know C compared to HTML5? I don't see how implementing even simple applications in HTML5/JavaScript is easier than in C/C++.

This sounds very much as such a statement. But what do I know.
And if you read what I said carefully, you will notice that I did not say it would be harder. Of course, if you want to debug stuff, oh well..

That's an interesting extra requirement you just added. Why is that a given?
It is not a requirement. Notice how I used SHOULD instead of MUST.

Or it's a way to hide most of the ugly stuff via abstraction and supply most of the same functionality using the browser/rendering engine.
And it is a way to add a huge number of security issues, if you mix native code and HTML. Remember the days of ActiveX, and how many exploits were centered around it?

If this is your main argument, I do have to wonder why you ventured into arguments about how development would be done and how porting to other platforms was unlikely.
Obviously, with limited CPU resources, people generally strive for efficiency. This is why you won't see graphic intensive games, encoders, decoders, etc. done in anything but C/C++ This is why Mono is not that widely used on Linux. This is why most desktop applications are written in C/C++ and not in Java or C# (although some are written in stuff like python and other scripting languages).

And even if you're right: how does this nullify the usefulness of HTML5?
HTML5 is very useful for what it has been intended for: WEB. For other things, it's the wrong way of doing stuff.
 

The Following User Says Thank You to Radu For This Useful Post:
shallimus's Avatar
shallimus is offline shallimus
2011-10-04 , 00:20
Posts: 568 | Thanked: 969 times | Joined on Dec 2009 @ Toronto
#62
Originally Posted by mikecomputing View Post
]I dont get why theey dont merge Tizzot with WebOS who already is heavily based on HTML5?

and another thing about tizzon is will Tizzot support Qt in core I really doubt. why would samsung/Intel have intrests in something that Nokia has intrests in. I know nomovoc will support Qt but something is really smoking here. Even if I agree HTML is they way to go in long run but I doubt it can fully replace lower level languages like C/C++

so this means we will be iinroduced to YET another API. Question is are the community willing to learn yeat another new API? isnt Linux fragmented enought with different API doing same thing? Seriously atm I have NO trust in those big companys. And Linux foundation!? well I dont know! after the mess with maemo/meego/moblin/webos and now Tizzot.

well well lets see whats happen. Personally I will follow the Mer project and hopefully give them some kind of help too... Atleast by testing weekly images of "MerHCE" then it maybe "merge" with tizzot in the long run who knows......

but atm. I have no trust in those company projects were it more looks like all of them trying to "pissing in theyr revirs" and make it looks like they care about FOSS.
I often enjoy your posts, but I don't always agree with them, even if I agree with the drift of them. However, this is one of the better posts I've read on here in some time - thanks.
__________________
tinfoilhat.dll: Trojan horse detected
Sailfish want list: calendar bugfixes, glanceable agenda, Swype or similar
Evolution continues (but we're still pre-Cambrian)
 
fasza2's Avatar
fasza2 is offline fasza2
2011-10-04 , 01:08
Posts: 187 | Thanked: 96 times | Joined on Sep 2010 @ London, UK
#63
Originally Posted by Radu View Post
WTF... we have at least 3 libraries that have ports on pretty much
Totally irrelevant example, or 'red herring' as you like to say. If you write an application in HTML5 or GTK, for example, the user should not see any difference. it is about lazy application developers who don't want to make an extra effort to assure more efficient code.
Comletely agree on that m8, though not the way you'd think. Yes programmers are lazy, but thats what drives them and technology forward. Let me explain why. All these interpreted languages(including C and C++) and tools(Qt, TK, GTK, modules etc.) came to life becuse of their 'lazyness'. You could also argue using interpreted languages vs. assemby. We all know the reasons there and I belive we have simillar reasons for this change as well.

Anyone knowlegeble would happen to know a bit more details about this HTML5 buisness? I mean how is it gonna work, will there be a lib that the OS could look at and execute the 'foreign' code or how? I don't mind even if someone just gives a link which explains it in more detail.
 
Radu is offline Radu
2011-10-04 , 05:16
Posts: 303 | Thanked: 146 times | Joined on Aug 2009
#64
Originally Posted by fasza2 View Post
Let me explain why. All these interpreted languages(including C and C++) and tools(Qt, TK, GTK, modules etc.) came to life becuse of their 'lazyness'. You could also argue using interpreted languages vs. assemby. We all know the reasons there and I belive we have simillar reasons for this change as well.
They say that laziness is the mother of 9 out of 10 inventions. And it is true. This being said, laziness should not be used as an excuse to do crappy work.
I guess you mean that C is 'interpreted' in the sense that it is not machine code. Yes, C is less efficient than hand optimized ASM, however:
1. Over the years the compilers are doing a pretty good job in optimizing the binary code produced from C programs.
2. There are many CPU families out there (X86, ARM, RISC, etc.) and each family has different members with extended instruction sets, etc. So it would be a nightmare to port application from, say, X86 to ARM.

This is one of the reasons why C has been developed: to be almost as fast as assembly, but portable and relatively efficient in terms of coding time.
 
jo21 is offline jo21
2011-10-04 , 05:34
Posts: 572 | Thanked: 259 times | Joined on Jan 2011
#65
Originally Posted by gerbick View Post
It even fails as a Flash replacement. HTML5 video gracefully fails to Flash. It doesn't do queries to build navigation. Sites that have moved to HTML5 + Audio are exposing their files due to lack of security - Pandora HTML5 version exposes the *.m4a files for easy download.

Not all browsers support Canvas - blame the software engineers for that and it's not exactly fully supported in all areas quite yet. I deploy Flash, it'll work everywhere.
i can get m4a streams of youtube, any flash site =)

so vurnerability is on flash.
 
ossipena's Avatar
ossipena is offline ossipena
2011-10-04 , 06:17
Posts: 3,159 | Thanked: 2,023 times | Joined on Feb 2008 @ Finland
#66
Originally Posted by Radu View Post
I don't think I understand this. A mobile application for Meego/Maemo or even Android running native code can do this probably thousands of time faster.
so an iteration calculation app is probably thousand times faster. what about an average fart app, are the farts 3000% faster and louder?
__________________
Want to know something?
K.I.S.S. approach:
wiki category:beginners. Browse it through and you'll be much wiser!
If the link doesn't help, just use
Google Custom Search
 
Radu is offline Radu
2011-10-04 , 06:53
Posts: 303 | Thanked: 146 times | Joined on Aug 2009
#67
As I was saying before, HTML5 has its uses. Playing movies, sounds,and displaying formated text are what it was designed for, so a fart app would be quite 'good' for HTML5. But most people need something more than that.
 
lma is offline lma
2011-10-04 , 07:06
Posts: 2,802 | Thanked: 4,491 times | Joined on Nov 2007
#68
Originally Posted by ossipena View Post
so an iteration calculation app is probably thousand times faster. what about an average fart app, are the farts 3000% faster and louder?
Probably not. But if you build a platform aiming at the average fart app, that's likely all you're going to get.
 

The Following 3 Users Say Thank You to lma For This Useful Post:
Joorin is offline Joorin
2011-10-04 , 19:21
Posts: 726 | Thanked: 345 times | Joined on Apr 2010 @ Sweden
#69
Originally Posted by Radu View Post
Those are tools that all the C programmers that do GUI stuff already have.
Yes? You have them, I have them and we both know how to use them as in writing the code, compiling it and debugging the result.

But for others, not as versed in GUI programming as you, having to do all this to get a simple application is a lot of work and includes learning some pretty complicated stuff. HTML5 lets them use an interactive environment for debugging (using, for example, FireBug) and it takes almost no time to have something up and running.

And it's still possible to build pretty complicated applications using HTML5.

It is not a requirement. Notice how I used SHOULD instead of MUST.
If you wanted to express "I'd like it" or "It'd be good" or something similar, I'd recommend writing that instead of using "should" since that, really, says that you expect it and see it as a requirement.

(I spend almost every day at work reading requirements and "should" is the one special word that really really signifies a requirement. Together with "shall".)

And it is a way to add a huge number of security issues, if you mix native code and HTML. Remember the days of ActiveX, and how many exploits were centered around it?
This is a valid point, to a certain degree. You can do much using only JavaScript and the browser supplied mechanics for, for example, HTTP requests but the security model is definitely important.

I do hope the HTML5 group, together with the browser developers, have learnt at least some lessons from earlier attempts.

Obviously, with limited CPU resources, people generally strive for efficiency. This is why you won't see graphic intensive games, encoders, decoders, etc. done in anything but C/C++
Yes? I've seen some pretty impressive pure HTML5/JavaScript action games already but if you want to do serious number crunching, HTML5/JavaScript is, obviously, not the best way.

But it will work in most cases, albeit slowly.

An interesting example from my reality is the technology used in some IPTV set-top boxes. Motorola is one example. There, a special HTML/rendering engine, or a pretty vanilla Mozilla, is used for all EPGs and the portal stuff. The box typically comes with 128MByte RAM and a 200-300MHZ ARM CPU which is even worse than our sexy N900. And the box runs Linux.

A browser based solution is still a good compromise between feature set and general knowledge needed for development.

This is why Mono is not that widely used on Linux.
Isn't it?

This is why most desktop applications are written in C/C++ and not in Java or C# (although some are written in stuff like python and other scripting languages).
Have you had a look at some of the Java/OpenGL games coming out lately? Pretty impressive, I'd say.

When it comes to C#, it sounds a bit like you aren't up to speed on all the application development that is done using it. I do it every day at work and the resulting (desktop) programs are, without a doubt, as useful and fast as any C++ application as long as you stay away from I/O intensive stuff since that's hit pretty hard by the GC if you're not really careful.

HTML5 is very useful for what it has been intended for: WEB. For other things, it's the wrong way of doing stuff.
You keep saying that and I can respect that as your opinion but apart from that I'd have to say you're wrong. It's another way of doing things.
 
TiagoTiago is offline TiagoTiago
2011-10-04 , 23:26
Posts: 1,522 | Thanked: 392 times | Joined on Jul 2010 @ São Paulo, Brazil
#70
I think RFC2119 is relevant to this conversation
__________________
NOKIA N900: IT'S OVER NINE THOUSAND


How to use HotspotShield in the N900
Google Voice from outside the US + Free international calls: LEARN HOW HERE!!!
 

The Following 2 Users Say Thank You to TiagoTiago For This Useful Post:
Reply
Page 7 of 8 « First 56 7 8

Tags
goodbye nokia, html != c/c++, meego is dead

« Previous Thread | Next Thread »

 
Forum Jump


All times are GMT. The time now is 11:48.

Contact Us - Home - Archive - Top