- "Weak" votes were already being cast long before the rating system
- They are very clear on purpose
- 5 steps is a far cry from 100
Also I like the idea of the "weak" votes.
More than 5 choices would start being too much. It would immediately start being a rating rather than an opinion. Less than 5 choices (ie. practically 3) feels a bit too little. It's like an all-or-nothing choice: Either you like/hate it a lot, or you don't care. No middle ground possible.
And as you say, people were already making "weak" votes in the old system. Now they would have an official way of saying so.
Since I agree too, I'll just go ahead and quote myself from the past, when I made this suggestion of a 2-dimensional poll:
While your suggestion definitely has merit, it doesn't really offer the "weak yes/no" answers as suggested. Also as laid out, it's a bit difficult to visualize what you are really answering to. Maybe just two separate questions with their own radio buttons could be better even from a GUI point of view.
Idea: How about a compromise: Have both questions at the same time.
This would actually be a bit akin to the previous entertainment/technical quality distinction in the voting. The "did you like the movie?" is asking for the entertainment value and "should it be published?" would be, sort of, asking for the technical quality (not so much about the technical quality of the movie in the same way as with the ratings, but technical quality in the sense that it meets the minimum standards for publication).
And while at it, make both questions have the 5 suggested answers (ie. yes, weak yes, etc.)
Yeah, I'm not even going to bother arguing over it. I just see "is this movie publishable" turning into "is this movie well optimized" since that's all tasvideos ever is any more.
I'm really getting tired of that claim. It's wrong for two reasons:
1) Tasvideos have always been about completing games as fast as possible with superhuman optimizations. From the very first SMB test runs I remember Bisqwit doing. There was never a time when speed was not the main goal. (The only thing which was different in the first years was that tools were not perfect and thus it was almost impossible to create frame-perfect movies. However, that doesn't mean the goals have changed.)
2) There have been many movies where completing the game as fast as possible is not the main goal because it doesn't make sense in the game in question.
"Should this be published?" is the correct question. "Did you like it?" is not. If someone makes a submission that doesn't even complete the game, doesn't attempt to do anything superhuman, but simply does something funny and he writes in his submission text "please publish my machinima video because there are no other places where I can publish it to the internet", would you vote yes if you find it a funny video?
If someone makes a submission of Mortal Kombat 2 where he states that his goal was not to complete the game as fast as possible but to show as many combo moves, finishes and glitches as possible, why would anyone vote "no" if the question is "should this be published"?
Then, 9 gigabytes in the cromfs archive containing... helpful files for watching the movie files for various consoles.
Is that the uncompressed or the compressed size? If it's the uncompressed size, then how large is that cromfs partition?
My tasvideos directory has 38 GB of avis. I have at least 99% of currently published ones there.
If you want to rate a movie, up pops a list of all the currently published movies you've ever rated, from 1 to n. Number 1 is your "favorite" movie. Number n is your "least favorite" movie. You then choose a ranking for the current movie, from 1 to n+1, resulting in a new, ordered list.
You would be surprised how easy it is to get into a situation with such lists where you later examine your list and see that you really think that movie A is better than movie B, movie B is better than movie C, but movie C is better than movie A, for different reasons.
If all three movies are in consecutive order in the list, then it's not such a big problem: Just choose one ordering which is the best compromise. However, you would also be surprised how easy it is to get into that situation but so that there will be something like 20 movies in-between those three, and deciding on a good ordering is a headache.
Otherwise I think it's a good idea. (Of course its only problem is that it would require more or less rewriting the entire rating system and dropping all existing ratings, which might not be a welcomed change.)
Xkeeper wrote:
Also, I strongly dislike the movement away from "Did you like this movie" to "should this movie be published".
I don't. "Did you like this movie" is asking the wrong question. Even if you don't like a movie, that doesn't necessarily mean it's not worth publishing.
Personally I find many of the published movies quite boring and I don't especially like them too much. However, most of them are worthy of publication: They are well-executed, and there are no good reasons why they shouldn't be published. Just because I find them a bit boring doesn't mean they aren't good.
Conversely: Even if you do like a movie, that doesn't necessarily mean it should be published. Maybe it's entertaining and fun to watch (to you), but could be clearly improved, and the author should make an improved version by removing the sloppy parts and mistakes. Maybe the submission breaks the rules. Maybe the submission is worse than an existing publication. There may be a ton of reasons why something should not be published even if you like it a lot.
(A no, weak no, meh, weak yes, yes type system would probably be best)
For what's worth, I think this suggestion has merit. More choices than in the old (and now also current) system, but shouldn't have the problems with the 0-10 voting system. If people want something published, then they can go ahead and give a "full" yes vote, but they won't be skewing the ratings, and other more moderate people will have more choices for their votes.
Cpadolf wrote:
klmz wrote:
Should people re-rate those "over-rated" movies?
If someone has rated a movie higher than they actually think that it is worth then that person should probably re-rate them, yes :)
How likely is it that someone who voted a straight 10/10 in the submissions forum just to see his favorite run published will now go back and re-rate it? Clearly such a person doesn't care about the rating system, so why should he bother?
It's a bit of a pity that there are a bunch of movies with probably a bit skewed ratings, but there's little that can be done about it, other than resetting their ratings completely, which would be a rather drastic measure...
Audio codecs like Speex can supposedly go even lower, but they won't be supported by players.
AFAIK those codecs work well only for human speech (they have been specifically developed for that) and very poorly for music (which is the reason why we still have mp3 and ogg).
A 1001-hour long video is going to be bigger than the Internet
How low can mpeg4 bitrates go?
If, for example, 8 kbps (ie. 1 kB/s) was used as bitrate, the video would take "only" 3.5 GB. (Of course the audio will probably double or triple that.)
Sometimes people want to color part of the text in their post with the same color as the background as an anti-spoiler measure (so that in order to see the text you have to select it with the mouse), similarly to how it's customary eg. in the tvtropes wiki.
The tvtropes wiki has a dedicated, easy-to-use syntax for this purpose. Here you have to carefully select a text color which matches the background color as well as possible (something made harder by the fact that the forum posts don't use a pure white background).
I was thinking: How hard would it be to add support for a dedicated anti-spoiler tag to the forum posting format? This would simply automatically color the text with the same color as the background.
But if you want to print unicode characters, don't you have to decide which encoding you are going to use to print them? After all "unicode characters" are just numerical values between 0 and some millions, or whatever. You just can't print "unicode characters" without deciding which encoding you are going to use to print them.
Personally I have always used size_t instead of whatever_container::size_type because it's easier to write. However, technically speaking doing that is not guaranteed to work because, AFAIK, the standard doesn't guarantee that the two types will always have the same size.
In practice I think 100% of compilers in any desktop or server system will have an identical type for both. There might be some really obscure embedded systems for which they aren't.
Having an iterator as member variable of a class which can be copied/assigned is indeed dangerous, unless you implement the copy constructor and assignment operator properly. That's because the automatic copy constructor/assignment operator the compiler generates will copy the iterator verbatim (there's no way it can know that you want it to point to the copied member data container).
Since it's a vector iterator, the solution is rather simple, as Bisqwit stated: Use an index rather than an iterator.
If you really need an iterator, and your class must be copyable/assignable, then there's no way around it: You'll have to write a proper copy constructor and assignment operator which set the iterator of the copy appropriately.
If your class does not need to be copyable/assignable, then it's STRONGLY suggested that you disallow both operations (by declaring them private). It will catch such errors at compile time.
You know Warp, when I wrote that, I was like "hahaha I bet Warp is going to be that guy who wants to confront me about this lolol". Do I know my audience.
I bet 90% of my posts show disagreement...
Anyways, I was not confronting you. I was expressing my honest opinion. I do find 99.99% of pop and rock music out there rather bland and uninteresting. In the end they all sound the same and don't stand out from the crowd in any way.
If all you like is techno and trance, this might not be for you, but if you do like actual music
Define "actual music". What is it in techno and trance that isn't "actual music"? (And don't go with the BS that they use synthesized instruments, as if that was in any way related to the definition of "music".)
Personally I found most of those Elton John songs you listed rather bland and uninteresting, if not even outright boring pop music, with nothing special going on.
"Don't go breaking my heart" had a bit of groove in it, but not so much as to make it really interesting. "Goodbye Yellow Brick Road" is a more known song which can be heard often in radio, etc, and rather ok, but IMO otherwise not particularly special either.
Personally I find most pop and rock music bland and boring. They all sound the same and they have nothing special. There are only few exceptions.
No I did NOT use anagram generators thank you very much.
If it's of any consolation, I programmed the anagram generator myself, rather than using an existing one.
I wanted to use the extensive 12dicts word list because I felt online generators are probably using reduced dictionaries (eg. lacking inflections, etc) and thus may not produce all possible cool results. I didn't find any suitable anagram generator program to compile and use, so I ended up writing my own. (Making it fast even for the longer anagrams is actually not completely trivial.)
Also some of the game names produced rather long lists of results (sometimes even on the thousands, even when limiting the number of words in the anagram), and searching for something which would work was not always easy.
So it's not like I didn't put some work into this idea. :)
