The Off Topic Programming Thread!

[CodieMorgan]

Originally Posted by Dygear

Thank you for making my point for me, so I did not have too. On a side note, I hate SVG (as it seems very unwieldy right now). I wonder if the spec is as bad as some of the output files I've seen thus far.

And that mean everyone is gonna buy a new phone as soon as it is released just because of HTML5?

I doubt that much.

Thing about the large percentage of people that will be on XHTML1.0 phones - especially the unlucky ones that use embedded browsers - not all of them have 3D or Wireless broadband - in in some counties - are very unlikely to update their browsers on the ones that support update downloads.


RSA for example - bandwidth is crap - and expensive - and is GPRS half the time.

Backwards compatibility is a standard and unspoken rule - particularly with CSS


- if It doesn't degrade well - alot of people are gonna miss out for some time.


U should at least have HTML1.1 fallback . Going HTML 5 is also pointless to force if u don't use alot of it's features.

EG: degradability by using <div> tags instead of <article><description><footer> etc..... or have a regex-replace if the browser string reports a non HTML 5 browser....


If someone who is as poor as me is excluded from the audience - they will look for another service, if they one they looking at now is absolutist will not cover at least 2 years backwards comparability.

[boothy]

Originally Posted by CodieMorgan

U should at least have HTML1.1 fallback . Going HTML 5 is also pointless to force if u don't use alot of it's features.

You've mentioned XHTML1.1 in two posts now, I hope you're not serving it as text/html and also waiting for XHTML2.0

[DarkTimes]

Originally Posted by E.Reiljans

http://blog.sucuri.net/2010/06/googl...ing-stats.html

Looks like totally free and open-source nginx has almost same market share as IIS.

Well I guess, but given how many web servers there are in the world, a 3% difference still tallies to tens of millions of servers, so not really that close. What's probably more interesting is how far behind everyone else seems to be...

[Dygear]

I had some great lines in this post that I just deleted (I typed it up, but did not bother posting it as it was pretty blunt. I'm just a little too tired to type right now, and smooth myself out as so not to offend. So, I'll say this, when I'm doing a personal project I shoot for the stars. I want to see how far the tech as come, and I want to see how far it can take me.

When I'm working, getting paid to produce code, I agree with everything you said ... to an extent.

-- Have a good night guys, I'm on my ambulance in about an hour or two, and I need a nap before I step foot on it and see my first patient.

[GeForz]

Originally Posted by DarkTimes

Well I guess, but given how many web servers there are in the world, a 3% difference still tallies to tens of millions of servers, so not really that close. What's probably more interesting is how far behind everyone else seems to be...

And stats for 1000 server don't mean anything for "tens of millions"

That's like saying "The coin landed on heads twice so it is 100% it lands on heads for 200 tosses"

[E.Reiljans]

Originally Posted by DarkTimes

What's probably more interesting is how far behind everyone else seems to be...

That's because all shared hosting providers use apache2 due to compatibility (for example, MediaWiki prefers Apache2 with mod_rewrite).
For most common tasks, nginx is still way better. Even when serving static content, for example - our firms' server connected to 100 mbit/s
dedicated line was giving out around 600-700 GiB per day with apache2, and around 1.0-1.1 TiB per day with nginx. With apache2, CPU was
the bottleneck, with nginx, only network connection was (it was using only 20% of CPU and 15 MB of RAM!)

[Dygear]

Originally Posted by E.Reiljans

Even when serving static content, for example - our firms' server connected to 100 mbit/s
dedicated line was giving out around 600-700 GiB per day with apache2, and around 1.0-1.1 TiB per day with nginx. With apache2, CPU was
the bottleneck, with nginx, only network connection was (it was using only 20% of CPU and 15 MB of RAM!)

That's a damn impressive stat! I've never heard of nginx, I'm going to have to look this up!

[misiek08]

@Dygear, better delete your post :P

You didn't hear about nginx? It's (like for me) best alternative for Apache. It support something like mod_rewrite and you can easy migrate from Apache to nginx setting nginx as proxy.

[Dygear]

Originally Posted by misiek08

@Dygear, better delete your post :P

You didn't hear about nginx? It's (like for me) best alternative for Apache. It support something like mod_rewrite and you can easy migrate from Apache to nginx setting nginx as proxy.

I've had my head in the sand for a while now when it comes to web server tech.

[misiek08]

I was just kidding about your post. I was using Apache for 3 years. Then friend told me about nginx..... and everything started to work faster and my servers are now bored because of free RAM and CPU.

[Dygear]

Originally Posted by misiek08

my servers are now bored because of free RAM and CPU.

Lol, that's awesome.

[DarkTimes]

Originally Posted by GeForz

And stats for 1000 server don't mean anything for "tens of millions"

That's like saying "The coin landed on heads twice so it is 100% it lands on heads for 200 tosses"

Not really, because two times is statistically insignificant, but a thousand is a pretty reasonable number. That's about the same number of people they ask in government polling.

[Dygear]

Anyone else using Stack Overflow for their programming questions? They have provided some very insightful answers.

[amp88]

Originally Posted by Dygear

Anyone else using Stack Overflow for their programming questions? They have provided some very insightful answers.

I hit it from regularly when I'm looking for help with problems but I haven't asked any questions there myself.

[Dygear]

Originally Posted by amp88

I hit it from regularly when I'm looking for help with problems but I haven't asked any questions there myself.

I've asked a few and all answers are spot on.

[misiek08]

I'm using it too.
There is no forum that can be more helpful than SO.

[PoVo]

Hey, I just wanted to ask a small question.

(This is all in C# @ LFS_External)

Is it possible to create a BTT button, so that when I click it, I get the text box, and when I click OK, I get another text box?

Long Sentence Short:

Click Button "Example" -> Textbox 1 shows up. -> Type in stuff, press OK -> Textbox 2 shows up. -> Enter stuff, press OK.

I don't think it's possible (in my eyes, but worth a try)

AFAIK: It would be possible if I knew the OK button's "ClickID"??

[CodieMorgan]

Originally Posted by boothy

You've mentioned XHTML1.1 in two posts now, I hope you're not serving it as text/html and also waiting for XHTML2.0

Not at all . XHTML2 is totally not my thing - I don't like it - Im interested in HTML5 (and current XHTML 1.1 'Strict')...

I tread all XHTML as XML - when parsing it!


Change of topic :

I created this using GDI to prototype some buffer matrixes using multiple renderbuffers. (vb.net)



Now I want to start the final product - and am having trouble decing eather to go DirectDraw9 clean - SlimDX, or any other - as my little project has peaked in GDI (cant do more without killing GDI framerate now )

Any suggestions of either any .Net friendly wrappers, or frameworks ( NOT XNA please... ) ? I am already testing some stuff in Axiom (Ogre/Mogre) .. but I need something lighter ... otherwise ill be doing Raw Direct2D and DirectDraw

[Dygear]

Has anyone played with the Go Programming Language? It's one of the more intresting things to come out of the Portland [s]trip[/s], programmers conference.


On a side note, Vic, can you please add <s>strike</s> to the BBCode list in the admin panel here? Thanks, bye!

[Woz]

Originally Posted by DarkTimes

I like TDD for some things, I've had genuine fist-in-the-air moments when a test passes, but I'm having issues with the fact that I am writing two or three times as much code as I was before, plus I sense a real increase in the complexity of some of my code.

I enjoy some things, writing the tests first creates some nicely decoupled code, and as I say I have had some really great moments when a a complex algorithm finally passes. I was writing some code the other day that converted base 10 numbers into base 36, and I found the tests very helpful while I grappled with my bad mathematics.

It will be very useful when working with dynamic languages, such as Python, when you don't have the safety-net of the compiler, but I'm not convinced that for the sorts of GUI apps I write, that I am really getting an equal return for the time-investment.

Anyway I'm still having fun and I plan to continue with it for the time being.

I got TDD into place early in the project I am working on which is now in the region of 500000 lines of code. Without our unit tests we would never keep it stable. TDD is great at finding all the little issues that occur in what you think are unrelated code.

You are right in that you create 2-3 times more code BUT the offset is less manual tests are required. So the time very quickly repays itself. We even took the time to write automatic reflection based test that check attribute tagging and validators are correctly hooked on objects etc.

You just have to stop fighting writing the test code, after a while you feel a little exposed when you write code without the associated unit tests to back it up.

I hate writing code without tests now as it means I have to perform the tests that would be automatic manually. Time better spent coding tbh.

[DarkTimes]

Has anyone tried the BBC JavaScript library? It looks pretty nice, especially the widgets that come with it. Not sure it has enough going to replace jQuery however, but it does look interesting.

[Dygear]

Originally Posted by DarkTimes

Has anyone tried the BBC JavaScript library? It looks pretty nice, especially the widgets that come with it. Not sure it has enough going to replace jQuery however, but it does look interesting.

I've never heard of it, I'll have a look.

[the_angry_angel]

Originally Posted by DarkTimes

Has anyone tried the BBC JavaScript library? It looks pretty nice, especially the widgets that come with it. Not sure it has enough going to replace jQuery however, but it does look interesting.

Yeah it's not too bad, but to be honest the possibility of it going away (if the BBC bin it) and the fact that I've been using jQuery for so long, both personally and at work, hasn't made me move to it.

I will admit I've been meaning to port one or two of the widgets to jQuery though. Never quite managed to find the time (as usual)

[CodieMorgan]

Originally Posted by PoVo

.... if I knew the OK button's "ClickID"??

I'm seriously late noticing this but.....

What the heck do you mean by ClickID?

You are working in C# right?

You are trying to track a C# button click, or a click else were ( btt)?

[Dygear]

Originally Posted by CodieMorgan

I'm seriously late noticing this but.....

What the heck do you mean by ClickID?

You are working in C# right?

You are trying to track a C# button click, or a click else were ( btt)?

ClickID is a struct memeber found in the IS_BFN (Button FunctioN), IS_BTN (BuTtoN), IS_BTC (BuTton Click) & IS_BTT (BuTton Type) structs. ClickID is used to track a button in these cases:

• ClickID can be used to update a button's text via the IS_BTN struct, where the ClickID is already known. However it should be noted that in this case, all contents of the packet is ignored if the button with that ClickID is already being displayed, only the Text member of the struct is read in the packet, and updated on screen. It's position will remain the same, as the position members are ignored also.
• ClickID in IS_BTC, will tell you what button was clicked by the client.
• ClickID in IS_BTT, will tell you what button was typed in by the client.
• ClickID in IS_BFN, will let InSim know what you wish to do with that button along with the SubT member of the Struct.

Originally Posted by PoVo

Click Button "Example" -> Textbox 1 shows up. -> Type in stuff, press OK -> Textbox 2 shows up. -> Enter stuff, press OK.

Yes.

Send a IS_BTN struct with the TypeIn set to the max length of the text you wish to receive up to a length of 95 charaters, set it's ClickID and send the packet. Note down somewhere what ClickID you set. When you get a IS_BTT packet, where it's ClickID is the same as the one you set, read the data and send another IS_BTN function like last time.

[Dygear]

I was looking at licenses recently, as some of you might of noticed, and I was looking up how Linus and Stallman did it, and I stumbled upon this:

Tech Talk: Linus Torvalds on git - Yeah, ok. I think I'll use git, taking a tenth of a second to go though the whole linux kernal source tree is pretty impressive. It should be more then powerful enough for us pokey devs on the PRISM project.

[bunder9999]

Originally Posted by Dygear

git

it's not that bad actually... i'm using it for a game source repo.

i used to use svn a couple years ago, and i wasn't really interested in learning another RCS, but once you configure it and use it a few times, it's pretty manageable. the only thing that bugged me was that i couldn't get it to send patch emails properly.

[dawesdust_12]

I love git. Yes I haven't used it for anything mission critical (I could have been just as well off without SCM, and kept 1 copy I worked off of), but it worked remarkably well. easy to get setup (once the concept was understood), and made things nice for sharing with the other person I was working on the project with.

[DarkTimes]

An excellent free book that teaches programming and the C# programming language.

http://www.robmiles.com/c-yellow-book/

[Dygear]

Originally Posted by DarkTimes

An excellent free book that teaches programming and the C# programming language.

http://www.robmiles.com/c-yellow-book/

I never really grokked, or even got C#. I felt that I was always on the cusp of being able to understand it. So, I'll read this after medic school is over