Do You Think Pandora's Power Will Lead To Relatively Lazy/sloppy/i


Status
Not open for further replies.
I have a tendency to only write software that has non-trivial performance requirements, so I can't really afford to write especially inefficient code.
 
People have worried about this every time technology got better. Developers, like all people, are supposed to be lazy. That means use the path of least resistance.

Just don't worry too much, if something is THAT slow, someone else can fix it.
 
benjymous said:
It's far better to install Dos/Win3.11 on a modern machine - boots in no time at all :)
Have win98 installed on a dual-boot laptop with 2ghz dothan and ati9700, 1.5GB . The beauty: even found drivers for all the hw. Had to steal a bit of RAM for a RAM disk before win98 boots, else 1,5GB would have been too much for it. Probably the fastest win98 one can have, totally love it. Feel safer when going online with thatone.

back on topic:
I can understand the points made, but regarding emulators, I would prefer they were heavily optimized or even written in assembler. Its always a plus when they can run well on an underclocked cpu (battery life).
 
Last edited by a moderator:
kanzlr said:
so, what Parkydr just said basically meant: he will port GravityForce2X to the pandora!
I won't, I can't justify the cost, but it would be an easy for someone else to port to the pandora.

I'd be inclined to go back to the original, most of the changes I did were to resize the view port and fonts - the pandora can display it using its original (better) resolution. It may well be playable without any changes using the keyboard.

When I say I'm lazy, I don't deliberately write inefficient code but I wouldn't start by writing some weird impenetrable code in case it happens to speed things up, I'd write the code then profile it to find the real bottlenecks then fix those.

It's yet another 80-20 rule - 20% of the code uses 80% of the processing so the lazy way is to fix the 20% that matters.
 
Last edited by a moderator:
Parkydr said:
I won't, I can't justify the cost, but it would be an easy for someone else to port to the pandora.
Dude, you should be on the free list! :ph34r:
 
Last edited by a moderator:
Yep, it's one of the core mantras in game development - premature optimisation is never a good idea. It's far better to have a complete system that you can profile, since you'll often be surprised as to which parts actually *need* optimisation.

You can waste months optimising a core bit of code and only gain a couple of frames per second, and the cost of making code that's far harder to debug later when things go wrong.
 
GunPei2X said:
Dude, you should be on the free list! :ph34r:
I'd love one but it's debatable how much use I could be on the Pandora, a lot of stuff will probably work with just a compile or may even be available ready packaged from existing repositories - any of the devs who have got a free one, or buy one could do that.
 
Last edited by a moderator:
This is sort of a silly thread, but perhaps its not obvious to non-coders.

People will code to the requirements and to their skill.

If you do a 'lazy port' (which everyone will try first, since its _easy and fast_), and its 'good enough' (for the porters requirements, be they his own or public glory or whatever) then he'll stick with it. But if he wants more (better performance, better UI, whatever) and he has the skill, he'll do that.

A lot of Linuxy apps have no UI (basic command line GUI) and other people supply shells say, or the shell is optional, etc. Certainly handhelds generally have very different UI needs than desktops, so the UI usually has to be redone -- you want it to fit usably onto small screen (regardless of rez, size counts), and also you want 'gamey UI' for the handheld - HUDs and such - that desktops may not care about.

Also, Pandora isn't as beefy as a desktop. Big fat emus that are already inefficient yet awesome (DOSbox, I'm looking at you) will need work anyway.

So anyway..

You can't "criticize" someone for taking the lazy approach. Laziness is a virtue ;) But you can criticize the end work. If you think something is no good, it needs work. But if a 5 minute port is really good, well, fine :)
Its usually not that easy either. Its a balance, since a 5min port might be 'good enough' for some, but not others.

--

But with the DNA of this scene, you can bet things will be pretty good. It is a bonus to have the option to take it easy.

Imagine - if you can have a whole pile of simple lazy ports in the first couple weeks. That is a win. Then, over the next few months, they get optimized and cleaned up. Also a win

jeff



EDIT: To help illustrate Jeff's point, I have gone back over his quick-and-dirty work and "optimized" it for spelling and grammar ;) -Chip
 
Thanks, Jeff.

Since I doubt anybody is going to improve on that explanation, we can put this (questionably useful) thread to rest.
 
Status
Not open for further replies.
Back
Top