Action button labels: poll 2 bis

Which proposal do you like best?


  • Total voters
    61

Too bad. Your opinion is not any more important than anyone else's, so the same restrictions apply to you as everyone else. Don't like how the poll is set up? Complain, don't vote, make your own poll. Don't consider your opinion so important it deserves breaking the rules to present.

EDIT: Yes, your opinion is being normalized by design.
 
Last edited by a moderator:
Too bad. Your opinion is not any more important than anyone else's, so the same restrictions apply to you as everyone else. Don't like how the poll is set up? Complain, don't vote, make your own poll. Don't consider your opinion so important it deserves breaking the rules to present.

EDIT: Yes, your opinion is being normalized by design.
The poll methodology then requires the voter to -lie- and state that unacceptable options can be ranked.

That is unacceptable. My vote stands as is.

To claim 'normalization' you must first have a semi-continuous classification - not disparate categories.

A better option would be to set it up as a Likert scale. http://en.wikipedia.org/wiki/Likert_scale

Allow each user to assign 0 to 5 points to EACH option individually. That would get the answer that this poll seems to want to obtain.

The current design effectively forces the voter to put the majority of their 'points' into categories that they -don't want-.
 
You're not assigning points. You're ordering five things.

This is five things: 1 2 3 4 5

This is five things ordered by preference: 3 2 1 5 4

This is five things ordered by preference: 2 1 5 3 4

This is NOT five things ordered by preference: 1 1 1 1 1

This is NOT five things ordered by preference: 1

EDIT: The scale you presented awards extremist voting and enables hostile voting behavior, where few negative people can undo a lot of generally positive ones. This is, in my opinion, not desirable. We are not looking to please a few extremists, we are looking for the most acceptable compromise.
 
Last edited by a moderator:
You're not assigning points. You're ordering five things.

This is five things: 1 2 3 4 5

This is five things ordered by preference: 3 2 1 5 4

This is five things ordered by preference: 2 1 5 3 4

This is NOT five things ordered by preference: 1 1 1 1 1

This is NOT five things ordered by preference: 1

EDIT: The scale you presented awards extremist voting and enables hostile voting behavior, where few negative people can undo a lot of generally positive ones. This is, in my opinion, not desirable. We are not looking to please a few extremists, we are looking for the most acceptable compromise.
And the MOST LIKELY outcome of this poll is going to be options that VERY FEW actually want as it will far over weight unacceptable options by forcing positive points to be assigned to them.
 
I'm sure I agree with your preference grench, but I think they have a point in what they're saying. Least favorite option would be the option you like the least. It would still be the same vote if your least favorite option were to be ceremonially thrown into a volcano. It's the least favorite option.

If they somehow count least preference into having some weight into what people want. (turning a negative view into a positive one) then they would be wrong. But saying I'm only voting for the options I care for and none that I don't is almost as wrong as only voting for one option for all votes, it screws with the results.

giving a value to the votes similar to the 2nd poll of +2/+1/+0/-1/-2 would still work here even if the wording is not right, it would work out exactly the same
 
Last edited by a moderator:
The question is to rank them from best to worst. That can mean you value them as +5 +4 +3 +2 +1, or as +2 +1 0 -1 -2, or in Grench' case maybe it is +100 -100.01 -100.02 -100.03 -100.04. This poll says nothing about the weights, only about the order. There is no assumption that you actually like your second choice, in fact maybe you don't even like your first choice. There is also no assumption that you dislike your fifth choice. The only assumption is that you like your i-th choice better than your j-th choice for all j > i. So if you are happy with something, then you should also be happy with everything you ranked higher; if you are unhappy with something, you should also be unhappy with anything ranked lower.

If you can't represent your preference as a total order (i.e. one of 5! possible permutations) , then you should either pick a total order that comes close to your partial order, or not participate at all.

This is why I wanted to put only 3 options in the poll, so there would be only 3! = 6 permutations and the poll could just be a single question with no possibility to cheat.

@Menesas: you put the Pandora layout in both the first place and the fourth place, and did not rank proposal 3. Please fix.
 
I have acquiesced and forced the answer to fit your design. May I suggest a weighting?

#1 answer 8 pts.

#2 answer 4 pts.

#3 answer 2 pts.

#4 answer 1 pt.

#5 answer 0 pt.
 
And the MOST LIKELY outcome of this poll is going to be options that VERY FEW actually want as it will far over weight unacceptable options by forcing positive points to be assigned to them.
You are not the one who decide how many people want it or not. This is a fair poll to determinate it, if you aren't agree with the poll, do not vote. If something is unfair, prove it. Instead of acting like an hysteric kid.
 
Last edited by a moderator:
The matter behind the poll choices is rather complex.

I think that polls are a severely misguided method of finding useful labels. Even if the polls are merely supposed to be used as a general overview of what people supposedly find acceptable or unacceptable, I think that they simply filter out way too many details. (Not to mention that the polls will most definitely inform the decision making process anyhow.)

On the other hand, the previous free-form discussions were meandering and rather unorderly.

Maybe a sensible next step could be to go ahead and aggregate the available findings in a concise manner, in order to provide it to all participants of the discussion as a clear set of things to be considered. What I mean by that is not to have a list of sugggested label combinations, each with "Pros" and "Cons" attached to it, but to have a (sort-of) definite list of these benefits and shortcomings, and of the entirety of other aspects that have to be considered.

I have not read all of the posts in the previous threads about this topic. I do not really have a clear enough picture of the whole matter at this time in order to make such a list, but I didn't want to just say "Too many polls" without suggesting something else. _wb_, do you think that it would make sense if you created such a list, and do you think that it would be a feasible task?
 
The only option I would vote for is 1) Pandora layout + 2 extra keys (A B Y X)

KISS
 
This discussion is so senseless... Did someone even ask EvilDragon, whether greek symbols even come into question?
 
Last edited by a moderator:
That is unacceptable. My vote stands as is.
You are taking this way too seriously.
The outcome is serious. 4 of the 5 options presented BREAK compatibility with existing Pandora applications. I consider that to be a serious flaw.
No, they don't. At least not in a serious way. Changing the screen size and resolution, changing the location of keyboard keys (e.g. the number row), adding two shoulder buttons and two action buttons, moving from softfp to hardfp, those all "break" compatibility in a way, but sometimes you have to break backwards compatibility to move forward. There are ways to mitigate compatibility issues.

If backwards compatibility were the only criterion, the Pyra should just be an exact copy of the Pandora, because anything that changes will reduce backwards compatibility. You can't even get a screen with a better response time, because e.g. my game NubNub does some ghosting compensation that probably turns out to be overcompensation on faster screens. Does that mean we should not aim for a better screen? Hell no! I'll just release an update to NubNub when the Pyra appears -- I'll want it to run at a different resolution than 800x480 too.

Changing the labels breaks backwards compatibility, but that's a very mild problem. As we know, people play emulators all the time. They can live with button labels that don't match. We can easily put a tutorial somewhere of what the button layout is of other consoles (Nintendo, Sega, PlayStation, Pandora) and how they map to the Pyra buttons. The real important thing for backwards compatibility is to keep at least 4 buttons, in a cross layout, mapped to Home/End/PgUp/PgDn. We have to keep the positions of those buttons like they are on the Pandora, otherwise games can become unplayable -- e.g. if a game uses the action buttons as a secondary d-pad, you don't want up to be left and right to be down all of a sudden.

---

My honest opinion about proposal 1:

The Pandora layout mildly sucks. I have learned to live with it, but it mildly sucks nonetheless.

It does not make sense -- A B left to right, OK; X Y bottom to top, huh?

It does not correspond to any existing layout (except GP2X).

Most of all, it does not take into account that there are already A B X Y buttons on the keyboard (a problem none of the dedicated game consoles have to deal with).

Adding two more buttons to that mildly sucking layout will only increase the suck. I mean, just look at this:

C  Y 

A      B

Z  X


It is a mess. It looks horrible. Spock and Sheldon Cooper would both get a heart attack from this. It has all the elegance of a x86 running a legacy COBOL program in some ancient PDP-8 emulator in Microsoft Windows ME. That is, none at all and please kill it with fire, please.

The only compelling reason to pick a horrible layout like that is backwards compatibility. As a rule of thumb, if backwards compatibility is the only good thing about something, it's time to break backwards compatibility. Especially if breaking it does not actually break anything (everything will still work fine).
 
Last edited by a moderator:
Seriously, these threads are good for two things from the perspective of the team:

1) Brainstorming out new ideas, in case there are things that the team hasn't thought of.

2) Keeping the board denizens occupied, so that the team can work in peace without us all going "Are we there yet? Are we there yet???"

At a distant end, the general gist of the conversations and the results of the numerous polls might be useful in case there are two design possibilities which are otherwise equal (or, which might actually be the case here, where the final decision is of no measurable importance whatsoever), but generally, the team will have to make all design decision based on all those factors that we don't have access to - Price, availability, intra-part compatibility, access to documentation, all that stuff.

And Grench, grow up. Yes, yes, we all know that you Really Don't Like any option that is not your favourite option. It is no secret. You have made that opinion perfectly clear. You really, really are going to throw all the toys out of the pram unless you get your way. You would really like to be able to score down things with a lot of points in order to drive home the importance of this toy-throwing.

Duly noted.

This is not a scoring poll, though. It is to establish a partial ordering , or maybe even a total ordering (which are really nice things with theoretical properties and all that, if you haven't run into them before. Honestly). It is a different kind of thing from a scoring poll. In an ordering poll, you do not score. In a scoring poll, you do not order. Different things. Different actions. OK? Can we stop with the whining now?
 
_wb_, do you think that it would make sense if you created such a list, and do you think that it would be a feasible task?
I could try to do that, yes. I think we have exhausted all the arguments pro and contra in the previous threads, so I could give a summary of that. Maybe tomorrow or so, it'll take some time.

In an ordering poll, you do not score. In a scoring poll, you do not order. Different things. Different actions. OK?

Sigh.

Grench just calmed down and changed his vote to something valid, now Thorgan shows up and votes 5 times for proposal 1. Is this poll so hard to grasp?
 
As we know, people play emulators all the time. They can live with button labels that don't match.
I'm not sure that's a fair comparison. People expect emulation to be different, and for the most part the labels are entirely irrelevant as you just map the emulated buttons to their expected positions. Old games and applications wouldn't be emulated, they would (hopefully, in theory) just run. How do you tell the difference between one game asking for Button X (pg-up) and another game asking for Button X (pg-dn)?
 
Back
Top