For more than 3 days I have been completely unable to download any videos. My torrent client is seeing half dozen of peers for each video, but nothing is downloaded. They have been all 0 kB for many days. Is anyone else having this problem?
Also, you should use !! instead of ! for mainlevel headers so that you can use ! for subheaders
IMO the logic should go the other way: The more the "header" characters, the "smaller" the header. In other words, 1 header character = main header, 2 header characters = subheader, 3 header characters = subsubheader, etc. That way it's much easier to create hierarchical sections (each with their own header) on the fly, while writing the text.
If the logic is in the other way, it takes a lot of work if you want to create a new level of subheaders but you already spent your "smallest" header level of one header character. You would have to go through the entire text and add a header character to each header.
In my experience if you want to make a rerecording emulator for some hardware, you will have to do it yourself (usually by patching an existing open-source emulator). Nobody else will do it for you.
Another question is whether TASing freeware games makes too much sense, unless the game is really popular. The problem with most freeware games, especially those for obscure hardware, is that almost nobody has ever played them. The good thing about commercial console games is that there usually is a very large amount of people who have played them and may be interested in seeing a TAS.
I'm not saying TI83 TASes shouldn't be published. I'm just raising a question.
Yes, this is definitely in the right direction. Could use some fine-tuning, though (for example it's not immediately obvious that you have to click on the number in order to get to the movie page).
Warp, you didn't answer my main question. Are you proposing that we keep the current system AND offer a mini-thumbnailed list as well?
Yes, my idea was that the mini-thumbnails page would be an *alternative* to the current Movies-* pages, shown by default, and perhaps with a link to the full list (something like "View full-sized list of movies (warning: large)".
Two possible ideas: Just put all the small thumbnails (with perhaps a small text attached to them) all in one line (letting the browser wrap the lines) or into some kind of table (eg. 6-8 thumbnails per line), or alternatively make a list similar to those in MovieStatistics, but with added thumbnails.
Are you saying that we should have the existing movie list (with full images) stay as it is, but add in a totally new page list with thumbnailed screenshots? Seems like a lot of effort on Bisqwit's part.
Do you think Bisqwit constructs the existing movie list page by hand? Of course he uses a PHP script which automatically constructs the page. It's not harder to make a small script which would construct the thumbnails-page.
Has anyone actually complained about this before? It seems like the only people who would struggle with this would be 56k modem users, in which case a thumb nailed image list of our extensive submission library would still kill them.
Finding a specific video for example from the Videos-NES page can be quite difficult. Besides, it takes a long time to download even with a regular ASDL.
Besides, making the site lighter and easier to navigate can always be done even if nobody has "complained" about anything. Not all features must be the direct consequence of complaints.
Who said anything about *replacing* anything? I said that this would be a navigation page to the individual movies, and that you could still get the full movie list (as given currently) if you want. It would just be a much lighter and easier way of finding what you are looking for.
Edit: Ok, I used the word "replace" in my previous post, but I didn't mean replace the current snapshot, I meant it like "in the vast majority of the traffic which happens when someone opens the Movies-NES and other such pages, the large images would be smaller ones.
You say reduced amount of traffic, but these are images we are talking about.
The current full list *is* full of images, one for each video. Relatively large ones. My idea was basically that these images would be replaced with much smaller ones (taking thus less space), as well as reducing the amount of text. Each image would be a link to the info about that movie.
Why do some still use lzh? It's one of the worst archivers which have been used widely. Perhaps they just are really old-school and they get retro-vibes from using that format?-)
(Here's one comparison I have made of several archivers, including lzh: http://warp.povusers.org/ArchiverComparison/ )
At one point in the very distant past the site used mini-thumbnails to make it easier to navigate to the desired video. You can see this here:
http://bisqwit.iki.fi/jutut/nesv-old/
This is actually a quite good idea, IMO. Current pages like http://tasvideos.org/Movies-NES.html are really long, take a lot of time to load and it's difficult to find quickly what you are looking for. Using mini-thumbnails instead could make this much easier and also reduce site traffic a lot (I'm quite sure many people who arrive at the site for the first time just click on that link, causing a whole lot of traffic).
The mini-thumbnails might not be exactly as in the old site, but reusing the idea in one way or another could be a good idea IMO. Both the visitors and Bisqwit would benefit (because of the reduced amount of traffic).
Edit: I don't mean that the current Movies-NES.html (and the others) would be *replaced* with the mini-thumbnails table, I just mean that it should be an alternative (and default) way of viewing those long lists. There could then perhaps also be a link to the full list be (perhaps with a warning about the size of that list).
Perhaps slightly off-topic, but:
I'm trying to run Pokemon Sapphire in the linux version of Visual Boy Advance 1.7.2, but I'm only getting a white screen and nothing else. I tried with some other game and it seemed to work ok. Does anyone have any idea what could be wrong?
Personally I agree with thegreginator. IMO a legit TAS is a run of a game which starts from scratch, ie. from reset and completes the game as fast as possible, even if some rational limitation is imposed (eg. no warp glitches or having to collect all items), solely with timed keypresses.
Cluttering of the "regular game section" could be alleviated by other means. It's not like "TAS" and "concept demo" were the only possible dividing categories.
How about this division:
1) Pure TAS: Completes the game as fast as possible, no matter what the means, "any%".
2) Restricted TAS: Completes the game as fast as possible, but imposing a rational restriction for entertainment purposes (such as using no warps or collecting all items).
3) Concept demos. (Most usually runs which start from a savegame instead of reset.)
which would involve running pattern matching algorithms on each line as it's rendered
Says who? The NES has a tile memory and a sprite memory. It only requires that when the game loads a graphic to either memory, the emulator internally uses a higher-version of that graphic when that tile/sprite is drawn on screen.
There are two options for replacing the tile/sprite graphics: Either examine its contents and search it in the emulator's stored database of graphic mappings (this is a quite fast operation), or examine where from the game code the graphics are being loaded and from this address make the deduction (an even faster operation).
In theory an emulator could even be enhanced with an "editing" capability for replacing graphics. At any given moment press some designated key which makes the emulator go to the editing mode. This mode pauses the game and then you can point to any tile or sprite with the mouse (the emulator could help finding them by drawing borders around them) and then tell the emulator what to replace it with (eg. from a list of existing higher-quality graphics read from a directory or whatever). When this is done, the emulator would then, depending on the replacement mode described above, either put the original tile graphic in its database, or resolve where that tile graphic was loaded from and put that address in its database, along with the higher-quality graphic.
This could actually go even further. The editing mode could support copying a tile/sprite graphic (or a group of them) to the clipboard, allowing you to edit it with your favorite image editor, and then it could support importing graphics from the clipboard as the replacement graphics.
Exactly in which country is bittorrent illegal? I don't know of any.
I said it implies you're doing something illegal, not that it's illegal itself.
"Implies"? In which context is this "implication" any relevant? You can't be fined for using bittorrent, even if it "implies" you are downloading something illegal. Heck, I'm sure that at least 90% of *internet* users download illegal material, thus using an internet connection equivalently implies you are doing something illegal. That's not enough to sentence you anywhere, though.
OK, but how do you detect what tile has been loaded?
You study the game binary and find the routine which loads tiles or sprites, and patch the emulator so that when that routine is being emulated, it takes the original graphic address used by the routine and maps it to the equivalent enhanced graphic instead.
(Of course an alternative "brute force" method is to take the original tile/sprite graphic itself, compare it to a list of graphics stored in the emulator, and then use the enhanced version for that graphic.)
It's not like this kind of thing hasn't been done before. I think Bisqwit once patched fceu to show a solid beam in megaman instead of a blinking one.
Isn't it possible to have the "nesvideos" text hidden from the viewer but still in the site's source code so that only the search engines see it?
Search engines (AFAIK) usually give a higher score to a text appearing in the title of the page than in the body. They might also do some html parsing and might even skip text not normally rendered.
...Ahh, I think it's moot; you could never expect a TAS for an existing emulator to stay in sync in the "enhanced" version, could you?
Why not? Why couldn't the emulator emulate the game *exactly* in the same way as with the original graphics, but simply replacing the game's graphics with enhanced ones. Everything else would be identical down to the smallest clock cycles.
creaothceann wrote:
A video console has certain standards imposed by the technology and the hardware. For example, the SNES can hold only 64 KB of graphics data in its main video memory. The registers used to access this memory have only certain bits reserved for addresses, so you couldn't access extra memory blocks even if there would be more. The main CPU and the support chips run at certain clock frequencies, so there would be no time to process/transfer more data. Not to mention that the custom hardware wouldn't understand things like alpha support etc.
I don't think you understood my proposal. The emulator would not touch the game itself in any way, and would still emulate the game exactly in the same way as always. The only difference is that when the game tells the "hardware" (which is emulated by the emulator) to draw a certain tile or sprite, the emulator instead draws the enhanced version of that tile/sprite graphic on screen. This will not affect the game itself in any way, only how it looks to the user.
Bob A wrote:
i think it would be easier to port the game.
Actually, I don't think that's true. If you port a game you basically have to write *everything* from scratch. Changing the game graphics in an emulator only requires you to do the graphics and the little piece of code which replaces then on the fly when drawing them on screen, and that's it.
Besides, the whole point was that existing TASes could be run with the enhanced emulator.
