Wikipedia:Reference desk/Archives/Computing/2008 May 19

= May 19 =

for speed, put XP or Windows 2000 on an old PC?? (pentium 3, 667 MHZ, 256 MB RAM)
so I have this old computer running windows 98, but 98 isn't stable enough for me, I normally would put windows 2000 on it, but I read that htis is SLOWER than windows xp:

http://home.comcast.net/~SupportCD/XPMyths.html

Myth - "Using Windows 2000 over Windows XP will improve performance"

Reality - Windows XP offers better performance than Windows 2000 so long as the recommended Windows XP requirements are met regardless of the age of the computer. With 128 MB of RAM Windows XP is superior to Windows 2000 and all older versions of Windows. This includes dramatically faster boot and resume times and highly responsive applications. Performance only gets better with additional resources, particularly when you run memory-intensive multimedia applications.

But if you look at their source, you see it's Microsoft! So, of course they're going to say that (at the time).... according to independent people, would Windows XP or Windows 2000 be faster on my ancient hardware? Thank you! —Preceding unsigned comment added by Beaterofapc (talk • contribs) 00:49, 19 May 2008 (UTC)


 * If you adjust Windows XP for best performance (System Properties > Advanced > Performance Settings) it will be very similar in performance to Windows 2000. If the computer is going to be connected to the internet, I strongly recommend Windows XP, as it is more secure. Your system is adequate for XP. --67.170.53.118 (talk) 01:08, 19 May 2008 (UTC)


 * With the above change they would be imperceptibly different; even with the default settings it probably would be fine. XP's theoretical performance improvements are in booting, and application launching (since it prefetches things and whatnot). However, the kernels and most of the underlying stuff is very similar between the two OSs, with XP adding various small things and making the GUI friendlier. I don't see any reason not to go for it. XP is newer as well, so its support end date is further out. 206.126.163.20 (talk) 03:27, 19 May 2008 (UTC)


 * Here is your complimentary nudge to use Linux, especially Ubuntu or its derivative, Xubuntu. Thank you for using the Reference Desk service, and have a good day. --antilivedT 04:04, 19 May 2008 (UTC)


 * You'd be amazed what an extra 256M will do for your XP performance. For such an old machine, people are virtually throwing away or selling RAM for VERY cheap. Good luck finding some, though. Graphics card performance matters to a lesser extent, and here again you'd be able to pick up an upgrade virtually from a junk yard. Old PC's are becoming a major landfill problem worldwide. Sandman30s (talk) 14:04, 19 May 2008 (UTC)

Maximum dimensions for image files
Hi, what are the maximum height and width for various image file formats? Are there any formats without restrictions? I'd like to know about PNG, JPEG and TIFF. Thanks. --Kjoonlee 00:53, 19 May 2008 (UTC)


 * The PNM format doesn't have any (defined) size limits, although software processing such files of course may be more limited. —Ilmari Karonen (talk) 01:39, 19 May 2008 (UTC)


 * PNG uses 4 bytes to store width, and 4 bytes to store height (4.3 million pixels square)
 * TIFF uses up to 4 bytes to store width, and up to 4 bytes to store height (4.3 million pixels square) (2 bytes each is more common)
 * I don't know about JPEG, but I've seen dubious websites that claim 32Kx32K. I think JPEG treats dimensions more weirdly than PNG or TIFF.
 * You may be more likely to encounter operating system limitations at very large sizes. I am not an authority on this. :-) --67.170.53.118 (talk) 01:52, 19 May 2008 (UTC)


 * (edit conflict) As for PNG, JPEG and TIFF, if I'm reading the specs right, JPEG seems to be limited to 65,535 by 65,535 pixels, while PNG and TIFF seem to allow up to 4,294,967,295 by 4,294,967,295 pixels. Good luck finding an image viewer able to display that 16-exbipixel image, though.  Or a hard disk to store it, for that matter.  —Ilmari Karonen (talk) 01:55, 19 May 2008 (UTC)


 * And it's hard to imagine an instance in which it wouldn't be more sensible to split it into multiple pieces and just have it pieced together on the fly. --98.217.8.46 (talk) 01:01, 20 May 2008 (UTC)

Can Windows Media Center record/pause RTSP streams?
Can Windows Media Center (in Windows XP Media Center Edition or Windows Vista Home Premium) record/pause videos streamed from the internet (specifically WMA/WMV streams over RTSP)? Thanks. --67.170.53.118 (talk) 01:02, 19 May 2008 (UTC) (Edit: I meant RTSP streams.)

How does the Chinese Wikipedia interconvert between different flavors of Chinese?
The Chinese Wikipedia allows articles to be rendered in several flavors of Chinese, including, among others, Taiwanese Chinese (in traditional Chinese characters) and Mainland Chinese (in simplified Chinese characters). The conversion involves more than just a one-for-one substitution of characters, because of the following complications:
 * 1. The mapping from simplified Chinese characters to traditional Chinese characters is NOT one-to-one. The same simplified Chinese character may correspond to several traditional Chinese characters, depending on the context.
 * 2. There are terminological differences among the different flavors of Chinese. Something called by one name in one flavor of Chinese may be called by a different name in another.
 * 3. Contributors add to the same article using both simplified and traditional Chinese characters.

How does the Chinese Wikipedia do it? —Preceding unsigned comment added by 72.78.237.18 (talk) 15:25, 19 May 2008 (UTC)


 * See Chinese_Wikipedia for a quick overview, and Automatic_conversion_between_simplified_and_traditional_Chinese for the more detailed one, as well as various appropriate links to the actual conversion tables and such. 206.126.163.20 (talk) 00:56, 20 May 2008 (UTC)

Cartoon Animation in Adobe/Macromedia Flash
Does anyone know any good websites (or books, if no known good sites exist) for learning the optimal way to create animation in Flash? (By animation, I mean character-based cartoon animation, not tweening.) I've done several searches, but the sites I've found are extremely basic. In particular, I am interested in three main aspects of Flash animation:


 * 1) Effective but efficient lip-synching (e.g. how many mouth shapes are necessary for convincing but not overly time-consuming speech animation)
 * 2) Optimal workflow techniques (e.g. to what degree should objects/characters being animation be broken into symbols, what's the best way to handle all the different symbols, etc)
 * 3) Recommended drawing techniques (this is less important than the other two; I'm just curious if people recommend anything other than a graphics pad).

Any advice on good resources for learning about these things would be most appreciated. Thanks! H. J. Hackenbacker (talk) 16:57, 19 May 2008 (UTC)

IE7 bug
I just found a nice little bug in Internet Explorer 7. If you have a dropdown list with only one option, clicking on this dropdown list twice crashes Internet Explorer. It is 100% reproducible. Adding another option solves the problem. Where do I go to report this bug to Microsoft? J I P | Talk 20:21, 19 May 2008 (UTC)
 * From a quick Google, is useful Nil Einne (talk) 20:42, 19 May 2008 (UTC)


 * I was also interested in reporting a different bug in Internet Explorer 7 that I had spent hours tracking down. That blog post Nil Einne linked to above looks promising, but the link in that blog goes to a Not Found page after you sign in. In the directory of connection forums, the closest match I found is one called "Internet Explorer Beta Feedback," but it appears to be for Internet Explorer 8 Public Beta. The directory page says that forum was created March 6, 2006, and the blog post above is dated March 24, 2006, so I guess that original feedback forum mentioned in the blog evolved to cover a newer version of Internet Explorer. It's not clear if it's still appropriate to submit IE7 bugs there, and the application process to join that forum requires you to provide your address and phone number, so I didn't complete the application at this time. --Bavi H (talk) 00:14, 20 May 2008 (UTC)
 * The general Connect gateway for IE is at http://connect.microsoft.com/ie Nanonic (talk) 19:53, 17 November 2013 (UTC)
 * (JIP, you might join that forum anyway and see if it's appropriate for IE7 bugs, or if your bug was already reported. Or, you might consider testing IE8 Beta, and if it has the same bug report it that way. I might do this myself later on for my bug too.) --Bavi H (talk) 00:38, 20 May 2008 (UTC)

Misc Desk
Is the Misc desk being vandalized or is it just me. When i see the page it looks like there is another page right on top of it before there was a picture of avril? anyone else seeing this? --Nick910 (talk) 20:24, 19 May 2008 (UTC)
 * I assume someone has hosed up one of the headers that's trancluded... I tried to fix it but it makes the links not work in my browser. Friday (talk) 20:25, 19 May 2008 (UTC)
 * This happened to all desks I'm pretty sure since it was indeed to one of the headers, take a look at the talk page if you haven't already for more details Nil Einne (talk) 20:42, 19 May 2008 (UTC)

Wordlist generator or commandline dictionary attack program
I'm trying to bruteforce a truecrypt password which I semi-forgot. I haven't calculated if it's feasible yet, but I haven't yet found a program which serves my need anyway. TrueCrypt supports a commandline mode, so all I need is a simple program which can bruteforce a commandline program until success. I've found programs which can probably do the command line part, trouble is, getting the wordlist. Most of them accept dictionary/wordlistss and at least one can even modify words in the dictionary. But surprisingly enough, none of them can use two or more words which is what I want. I also want to specify that one word is always there. I've come across some wordlist generators, but most of them only allow you to specify stuff like R?????p? or someething of that sort. Anyone know of either a wordlist generator, or a command line bruteforce program that can try to bruteforce two or more words (specified of course) and always include at least one specific word? I appreciate of course that it's a power function so with several words it can quickly get into taking many, many years. Nil Einne (talk) 21:00, 19 May 2008 (UTC)
 * If the tools you already have are sufficiently Unix-y &mdash; that is, they use standard input and output &mdash; you should be able to just use paste (with the -d ''  option) and yes (and perhaps process substitution out of bash or so) to get a list of words with your word prepended or appended. Is that all you need?  --Tardis (talk) 23:02, 19 May 2008 (UTC)
 * Something like this will do it for two words:

check { truecrypt --decrypt-with-password "$@" && echo "Password is $@"; } known_word="celadon" for i in `cat /usr/share/dict/words`; do   check "$known_word$i" && break check "$i$known_word" && break done
 * --Sean 13:34, 20 May 2008 (UTC)

Slow internet problem?
i'm currently having a problem with my internet, webpages take an extremely long time to load, if it all, yet i can still view Youtube videos easily, with the video loading quickly, and i was able to download from Rapidshare at 90 kb/s. Also, this only happened after i messed about a bit with my internet settings to try and forward my ports properly for Soulseek. I'm using Mozilla Firefox and Windows Vista. Does anyone have any idea what i could do to try and get my internet web-page loading-speed back to normal? 91.109.10.126 (talk) 21:09, 19 May 2008 (UTC) —Preceding unsigned comment added by 91.109.10.126 (talk) 21:08, 19 May 2008 (UTC)
 * It probably won't work, but you could undo what you did and see if you still have the problem. --WikiSlasher (talk) 10:00, 20 May 2008 (UTC)
 * Does it happen in IE (ie. Firefox issues)? Using any security software? Router? If you surf by IP address does it work quickly (ie. DNS issues)? There are a lot of different points of slowdown... 206.126.163.20 (talk) 22:00, 20 May 2008 (UTC)

Visible polygon vertices algorithm?
I would like an (efficient, easy to implement, floating-point-tolerant, pony-based) algorithm for identifying which vertices of a set of polygons are visible from a given point (that is, not obscured by the interior of any of the polygons). The polygons can certainly have CCW-ordered vertices (which allows the useful inclusion of an inside-out polygon that contains the whole field), and are simple, but may be non-convex, have a large number of vertices, and/or be large in number. I have an algorithm that (I think) works in theory, but it's falling afoul of roundoff error. I'll give it, but I suspect that the true answer may be unrelated. The problems, of course, lie in calculating the intersections with the triangles and angles (or sectors): for a simple triangle with one visible edge, that edge attempts to hide the triangle from further consideration by resolving the interval of angle containing it, but the triangle may be later considered to just extend into the allowed region of angle. (It's also possible for the polygon whose edge defines a triangle to search to be counted as inside that triangle, although this can be worked around.) Any thoughts? --Tardis (talk) 23:45, 19 May 2008 (UTC)
 * Given: a list of polygons (each a list of vertices) and a query point not in any of them.
 * At all times, maintain a list of intervals on the torus [0,2π]; each interval is a set of directions in which we have not found the termination of sightlines. This list is initially the entire interval.
 * Also maintain a stack and a queue of edges of polygons, each initially empty.
 * Also maintain a set of vertices to return, initially empty. (A set because we will "notice" some vertices more than once.)
 * 1) Find the nearest polygon that (at least partially) lies in an unresolved direction and, if the edge stack is not empty, which intersects the triangle defined by the top edge and the query point. (This can be done efficiently with some sort of quadtree.)
 * 2) If there is no such polygon:
 * 3) Pop the edge stack. (If it was empty, the algorithm is finished; return the set's members.)
 * 4) For each endpoint of the popped edge, add it to the return set if it lies in an unresolved interval.
 * 5) Then mark the interval of angle subtended by the popped edge as resolved.
 * 6) If the edge stack is now empty, pop an edge from the queue and add it.
 * 7) Go back to the start.
 * 8) Find the nearest edge on that polygon (also efficient with an appropriate quadtree); if two edges are closest because the closest point (to the query point) on the perimeter is a vertex, and the query point is on the "inside" of one of those two (it cannot be inside both, or it would be inside the polygon), choose the other edge.
 * 9) Push the chosen edge on the stack.
 * 10) Walk around the polygon in each direction, pushing other edges onto the queue. Stop when the query point is "inside" an edge and do not push that edge.


 * This sounds like a subset of a hidden line removal program I once wrote. One part was to create a piercing vector from the point in question to the camera point and then check to see if it interesected any surfaces between the two points.  It sounds like you only have a wireframe model, though, not a surfaced model.  You may need to first create a surface model to do what you want.  Creating planar surfaces inside a polygon is relatively straightforward, I believe, and creating planar surfaces outside those polygons shouldn't be difficult, either.  You first must define the plane using any 3 non-colinear points, then apply boundary edges to them. StuRat (talk) 02:32, 20 May 2008 (UTC)
 * If I'm not mistaken, you're talking about a 3D problem, perhaps populated by flat surfaces; I'm talking about a plane problem (all polygons and the query point lie in it). Sorry for not being clear.  --Tardis (talk) 04:15, 20 May 2008 (UTC)