Poll: what if Unity dropped support for ancient GPUs? (2013 version)

If its a massive pain to continue supporting them then you may aswell drop it. You don't buy a super nintendo to play halo 4 The earliest model i've tested my projects on would have to be 2005. I wouldn't say 10+ year old GPU dont deserve games though.

 

It really depends on the amount of work on your end, if it is easy to keep supporting them then why not?
EDIT:
Oh I just read DX9, yes I would get rid of pre DX9 cards.

 

Yes if it is holding back Unity then get rid of them ASAP.

 

"prevents us from adding new fancier stuff. "
Keep in mind, this is the stuff, people really want new fancier stuff

 

Yes, but only if you do display a big pop-up message saying "Get a GPU from this millennium, dude" when there gpu is not supported

 

Pre-DX9? Holy crap, I didn't know you had archaeologists employed to even find specs for them

 

Can any GPU that old even run the graphical side of unity?

 

I happen to be a big proponent of that very sort of stuff

 

drop it. it will help the hardware industry

 

If it can't do much then why can't you add the new features and just make them not work on old hardware?

 

IMHO support "anything with pixel shader 2.0 and up". If you drop support for cards at mobile-level performance on the grounds of performance then maybe drop mobile too eh?

3% of how many people? if 3% turns out to be a million and your customers are doing web games its a risky move.

I hope you're not thinking of dropping support for vertex lit because many still will use that for pure speed, even on big games. Maybe especially on?

 

What about all those Apple Machines? IIRC Unity uses OpenGL on those machines. And IIRC, Apple hasn't updated it's OpenGL port since quite some years?

Wouldn't dropping support for Pre-DX9 cards also have an influence on Apple machines?

 

Is it possible to show numbers instead of percentages? I mean 1 million people representing 0.1% would be too good

 

No way, I don't want new fancier stuff. What's wrong with you!?

On a more relevant note..I wouldn't care if you dropped support for cards that old. Doubt the majority would.

 

.................

okay get rid of those old graphic card........

 

you could always use an old version of unity to work with those old things, if necessary

 

+1

 

QFT

Drop them.

 

Would a change like this have any impact on the mobile platforms (ios/android/etc)?

 

They're not using ancient GPUs, they use modern GPUs.

--Eric

 

Yes, stop posting here and go drop that support now : p

 

If 3% is a million, then you still have 33 million customers to target. Do you really want to spend that extra time servicing the 3% at the expense of creating truly compelling features for the 97% that would probably increase your sales?

Fold in the fact that people targeting older system can't use the new features anyhow, and that they can still use an old version of the software to keep support, and I just can't see it making sense to worry about the 3% of your customers that are holding things back and making development more expensive.

 

Wait.... are you saying using OpenGL ES 2.0 is actually slower than OpenGL ES 1.1 pre-Unity 4.2?!

 

I believe this resembles the IE-problem for web developers. If so, let them crash and burn. I know from experience how much needless work is wasted on this kind of stuff. I assume real development suffers a bit because of this.

 

In our experience, it depends on the situation. It has been a significant free performance boost if all you're using is fixed function on OpenGL ES 2.0 capable devices.

 

More than happy with Unity dropping support for pre-DX9 GPUs, and OpenGL ES 1.1 on mobile.

 

Exactly what I was going to say. If people are focusing on casual users playing low-fi web games they aren't going to really need newer versions of Unity anyway, especially since Unity will be dropping Flash as well. Drop pre-DX9 and put all those devs to work on a new terrain engine.

 

From your side what are the benefits of dropping the support? Is it extra development manpower made available or is there software benefits like smaller compile size or other such benefits?

Is it worth dropping the support of that 3% to better support the other 97%?

 

I say do more than drop them, throw them to the ground and step on them. Seriously, supporting these things is certainly a time sucker, I'd wager much more is lost supporting them than gained. And, it's no fun at all. You've gone beyond the call of duty supporting them as long as you have, IMO. +1 for the drop vote.

 

Keep in mind that 3% number can be very deceptive, since it's at the lowest end of hardware capabilities as well as customer willingness to pay for things. I wouldn't be surprised if most devs making web games with Unity aren't making content that will run effectively on those platforms, so even if that customer was just as 'valuable' disregarding their hardware, they're likely to lean towards simpler tech (flash games) because of the poor experience they'd get with Unity games.

I'd wager what is 'lost' would be WAY under 3%, and certainly supporting old APIs/hardware can have nasty influence on your codebase. You inevitably end up with weird, touchy hacks that frequently suck away small amounts of time just for maintenance. These costs are hard to quantify, but I've found them to be significant.

 

This could also be 3% that have installed unity webplayer i'm guess as they got stats from it but how many actually us it. How many pluggins have people installed over the years but never use them. So that 3% might not be important enough to support.

 

I actually have a 2002 Toshiba Portege M200 Tablet with Wacom Digitizer that still works well with Windows 7. That said I've never considered developing with Unity on it although it does still have use as a Gimp/Blender/OneNote/Office computer if I wanted to dust it off.

I would look at browsers traffic to make your decision and if you can ascertain the customers from that ask them directly. A colleague says he still repairs many 10 year old plus computers and that those folk have only ever used computers mostly as an internet browsing machine.

 

I actually WOULD mind if you did NOT drop them, i want less legacy and more high end, don't forget those of us that only care about DX11 ! (i'm not saying you should move away from DX9, but honestly you can't do everything, so no point to push DX11 and new tech on one side, and support 10+ year old cards on the other, i don't think the guy with a >10 year old graphic card is very likely to crawl the web for unity powered low end games that target PC and check that indeed those are compatible with his old hardware unlike everything else he tried)
I'd even be pretty happy if you dropped everything >5 year old whenever a new major release comes (4X 5X etc), that way it doesn't hinder development and those wishing to target lower end can stick to the previous main branch.
You could do like microsoft too, sell extended support to those who want legacy stuff supported (so say if you had pushed 4.X with only 5 years old card support, you could've sold additional 3.X support licenses that guarantee bug fix for X duration and that do support legacy cards)

 

I got rid of my Android arm6 with vfp because Unity 4 didn't support it. I thought you already dropped arm6 / openGL ES 1.1

I think if Unity is going to go for government / big business contracts then they need to get everything to openGL 2.0 / arm7 minimal target support otherwise you will be stuck with openGL 1.1 / arm6 for a long, long time.

 

Is anyone who has mentioned defense of the 3% being too many people, stopped for a moment consider the simple obvious fact that it is ONLY 3%!

 

Except that percentages aren't the whole story. Sure 3% of a 1 million installs is only 30k users, but according to Unity's figures there are 200 million web player installs, so that's actually 6 million users. If you happen to be a company targeting that low end with Unity, that's going to be a huge chunk of your market gone.

Honestly i'm in two minds about this. On the one hand it enables Unity to be used to target low end systems, which can be essential for certain clients. Whilst its also important to remember Unity is not simply used for games or game developers, there is a very wide range of uses for it across many fields. I disagree with the statements that these developers should just stick with old versions as it forces them into a dead-end should they require new features or discover bugs.

However on the other hand I can fully appreciate the chore it must be for UT to maintain this support for legacy systems, how it could make the code base fragmented filled with workarounds to get it to support such old tech. Its also very appealing to think that dropping such legacy support would enable UT to focus on improving and updating graphical features.

I think on balance, the lack of those wanting to keep pre-dx9 support in the poll and the probability that few developers do rely on such support (at least according to the replies in this thread) it does make sense to drop legacy support in the 4.x cycle.

 

Just remember that Unity makes downloads of previous versions readily available, and if you're a Pro owner, a license for the current version also covers all previous versions. In other words, dropping support for old hardware doesn't mean you have to stop using Unity.

--Eric

 

That's flawed logic, just because a game might be targeted at the low end that doesn't mean that it'll only be played on low end outdated systems. I suspect that the proportion of such outdated systems even on any low target games (are there any?) is still likely to be close to that 3%, in fact that's the only segment where the proportion could be that high. Losing the 6 million low end pc users will therefore only at most lose 3% of your potential market share, while anyone targeting anything remotely higher than the lowest of the low will actually experience no change at all.

All in all that means that e.g if generously speaking 10% of Unity games monetized were designed to target the lowest spec computers then for unity developers as a whole you'd actually see a loss of only 0.3%, then take into account the sort of person that's likely to play games and what their system would be, subtract from that the subset that's likely to monetize in any way and you're really looking at a much smaller proportion of potential users.

I think it's so negligible as to be a net drag on development costs and overheads to continue to support, by UT continuing to support this it actually slows down other developments and technologies which might expedite your own workflows and therefore you're potentially losing more money from UT maintaining old code than you're gaining through monetization of ancient pc users, that's not even taking into account your own time spent on the support.

 

I doubt that new features could work efficiently in such old hardware. As said before, older versions of unity are available for those wanting to provide support on older machines. I like the idea of moving forward.

 

Here's the thing, whilst Unity is undoubtedly game-centric, there are many developers who do NOT make games with it ( most of my work is nothing to do with games), treating it as such skews the logic that they can be ignored. There will be developers who create content for clients that do not need anything more fancy than dx7 level hardware and they will be forced to support that by their clients. It is for these developers for whom 6 million users can and will be their essential market share.

So it is not flawed logic in the least, just far too many developers here think that games are the be all and end all of Unity and that naturally reflects their opinion as to the importance or not for retaining legacy support. To simply dismiss 6 million users as of being no interest to them is self serving, though understandable.

That is my point about dismissing 3% as being a low number, as from experience I know of developers who have had to support those users for their clients in non-game products and losing those users is not an option. I didn't want that fact to be lost in the rush to drop the legacy support.

However as I went on to say, the lack of response to this thread by developers who continue to need such legacy support suggests that it can be dropped, that the number of developers it is likely to affect has reached a point where it is outweighed by the advantages to Unity and the Unity community as a whole. Though of course there is also every chance that those developers don't read the gossip section, so it will be a rude awaking for them when they do find out.

 

Exactly. Removing support doesn't affect me - I don't need to care about that 3% but many developers do. Another situation would be for unity to ensure the legacy platform continues to be compatible with new systems and browsers (within sane reason) so they can drop support - yes but should you want that support you use version x. This only works out if unity is willing to go back from time to time to ensure it still runs.

 

Support for old hardware should probably not be removed in a 4.x release. Classic version-numbering always ensures revisions remain compatible, and removes support/changes APIs only in major releases (unless it's a result of a major bug fix - but then the first number should probably be bumped up too).

Or to put it like this: The paying customers paid for 4.x, and it's within reason to expect it to continue to work for the same combinations of hardware until it's time for 5.x.

(The thought will probably make the Unity devs cry, though

 

They could escape this misery by jumping from 4.2 to 5.0

 

Historically, Unity's .x releases have often been "big" releases with major changes, so a .x release is fine for dropping support. They've done it before (e.g., dropping PPC support in 3.5). The .x.x releases are the minor versions, and I'm sure they wouldn't make big changes in those cases.

--Eric