Wikipedia:Reference desk/Archives/Computing/2007 November 12

Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

November 12

Split M:TG cards: a database trilemma

Most Magic: The Gathering cards have one name, one mana cost, and so on. But some, such as "Dead // Gone", have two, so in my PostgreSQL database I need to be able to split each over two table rows one way or another. There are several ways I could handle split cards, and each has pros and cons.

1. One option would be to split every card into a "parent card" and one or more (usually only one) "subcards." The parent card would hold the stuff that's the same for all subcards: name, set, rarity and number of copies. Each subcard would have an FK linking back to its parent card. This would probably be the most elegant solution and require the least disk space, but it would make form-based input difficult. The main card form would need to contain columns of two different tables, and each new/deleted card would result in a new/deleted row in two different tables. (Then there'd probably be a second form for second and further subcards.) Are there any free cross-platform front ends that can even handle that kind of form?
   There's also the issue of how not to make it look on certain views like I have twice as many copies of a split card as I actually do. Unless there's some way to do colspan in a view's output, I'd have to assign the number of copies to one of the subcards arbitrarily. (Dividing the number across all subcards would make it harder to read and require a switch from integer to floating point.)
2. A second option would be one table for main cards and a second table for only second subcards. But this would mean any time the table needed redesigning, I'd have to do everything twice. And if the two tables should go out of sync in structure, the combined view would look awful and might place data misleadingly in the wrong columns.
3. A third option would be to put everything on the one table, give all subcards the same set etc., and create a table linking pairs of subcards that were on the same card, with one subcard on each card being the master. But this could make formatting of views rather difficult.

What are the reference deskers' opinions on the best option? Can they propose other options than these three? BTW, this is going up on Sourceforge under GPL once I get my own data entered in and a few details worked out, so you'll be helping build a database for all MTGers, not just for me. —Preceding unsigned comment added by NeonMerlin (talk • contribs) 01:30, 12 November 2007 (UTC)

Are they just separate cards with similar data? I'd say just enter them separately --ffroth 01:54, 12 November 2007 (UTC)

It sounds like it is one card with two different mana costs. If you know that there would be AT MOST only two different costs, I suppose you could just add another field (ie "Mana Cost 2" or something similar that would only be filled in for those cards (and left blank for others). If some have two, three, or four different costs (variable number of costs, something I haven't really seen in MTH, but I haven't played in a while), then a junction table is probably the way to go (sounds like this is what you meant in #1) --Bennybp 03:27, 12 November 2007 (UTC)

Dead // Gone looks to me like two cards on one cardstock. Weird. If two is the max per cardstock, I'd probably create an extra set of fields in each row rather than try to coordinate multiple tables. Coordinating multiple tables can work out well in practice but if the number of options is limited to two you might as well just hard-code it in; it'll be easier from a data-entry point of view and faster to search/use/index. That is, I'd just add two fields, the presence of non-null data in either would indicate that the card is one of these weird double-cards. So if your title and cost fields are CardTitle and ManaCost, add a CardTitleDouble and ManaCostDouble which default to null. --24.147.86.187 04:37, 12 November 2007 (UTC)

Wow, I stopped playing M:TG after buying Tempest... cards are weird now, apparently. For this specific card, I would treat it as if it were an Instant with two effects, one with activation cost of R, and one with activation cost of 2R. Imagine it is an artifact with two effects with different mana costs. -- ReyBrujo 04:44, 12 November 2007 (UTC)

No no, these solutions won't do at all. First, it has already happened once and could easily happen again that there are more than two spells on a card. Second, putting mana cost in two columns would make it impossible to sort by and make the card appear in both positions. The subcards have to be separate rows, that much is decided. NeonMerlin 07:10, 12 November 2007 (UTC)

My suggestion would be to go for (at least) two tables: one "card" table whose records each contain data about a single physical card — such as its name, the set it belongs is, perhaps a picture, etc. — and a second "effect" table describing what the card does, including things like the mana cost of each effect. That way, you can easily have cards with multiple effects. Not having ever actually played MtG, I'm not sure which table should contain the card type; if there are, or might some day be, cards with multiple types, then it should go in the effect table too. You might also want to split card data further into more subtables if, say, a card might be part of more than one set. —Ilmari Karonen (talk) 08:48, 12 November 2007 (UTC)

OK, but if you want our input on how to organize the data, you have to be a little bit more forthcoming about what the specifics of the data to be sorted are. One example is probably not enough if you want people to suddenly be acquainted to the exceptions of a system. --24.147.86.187 16:57, 12 November 2007 (UTC)

The real question is: do these "mini cards" exist in the game? If not, they can be considered abilities of the card, and be treated as such. Creating a "card" that contains other "cards" is a waste of time. Then you will need to add another field to mark which one are "true" cards and which ones are "mini" to prevent getting the mini listed as if they were real cards. That is, if the real cards do not exist by themselves. -- ReyBrujo 17:33, 12 November 2007 (UTC)

Yes. the mini cards exist. While a split card is in the hand, it has two mana costs, two types and all the subtypes and colours of both subcards. (This can cause it to make things happen twice.) While in play or on the stack, it is whichever subcard the player chose it to be when playing it. NeonMerlin 00:09, 13 November 2007 (UTC)

The standard approach to this is to separate out anything which you can have multiples of into a separate table. So you'll have a cards table with all the information which is unique to each card. Then there will be a separate spells table with spells and each spell has a field of card_id indicating which card the spell belongs to. Then you can get at everything quite simply using some simple SQL joins. Take a look at database normalization. Donald Hosek 22:02, 15 November 2007 (UTC)

my computer is blue screen on death

how do i fix —Preceding unsigned comment added by 172.168.247.148 (talk) 02:14, 12 November 2007 (UTC)

Could also be bad hardware. -- JSBillings 11:15, 12 November 2007 (UTC)

Or bad software. --antilived^{T | C | G} 11:22, 12 November 2007 (UTC)

You haven't given nearly enough information for anyone to answer your question. You might as well have said "my computer isn't working, how do i fix". You need to look at the information displayed on the blue screen, at the very least the main error code and the name of the offending file (if any). For example, in this image the code is PAGE_FAULT_IN_NONPAGED_AREA and the file is SPCMDCON.SYS. Or you could just update everything or switch to Linux. The latter would completely eliminate blue screens of death, because in Linux they're called "kernel panics" and aren't blue. -- BenRG 14:19, 12 November 2007 (UTC)

Out of curiousity, How are you typing your message now? XD

♥ECH3LON♥ 02:17, 13 November 2007 (UTC)

CHANGE THE WAY USER LOGS ON AND LOGS OFF

PLEASE HELP! WHENEVER I CHOOSE TO USE THE WELCOME SCREEN A BORING MSG. APPEARS SAYING "CLIENT SERVICES FOR NETWARE HAS DISABLED THE WELCOME SCREEN AND FAST USER SWITCHING.TO RESTORE THIS FEATURE,YOU MUST UNINSTALL CLIENT SERVICES FOR NETWARE" I CAN'T FIND THIS FILE(CLIENT SERVICES FOR NETWARE). WHAT SHALL I DO? —Preceding unsigned comment added by 117.99.30.197 (talk) 13:16, 12 November 2007 (UTC)

To uninstall the NetWare client, perform these steps:

Go to Start>> Control Panel
Select Network and Internet Connections
Click Network Connections
Right click on Properties and select the Network/Internet connection you use
Check each one individually by right clicking and selecting Properties
When you see the client services on Netware in the list of properties, click on it to remove the small 'X' Click nothing else at this point.
Now, scroll down and select Uninstall
Your PC should reboot with the old Netware logon. After it does, go back to Control Panel and click User Accounts and change how users logon.
You should now be able to re-establish the Welcome screen that Netware disabled.

Andy^reply 15:46, 12 November 2007 (UTC)

I am experiencing a major problem with Firefox

I hope someone can help me with this. Whenever, I edit any big pages on Wikipedia, a message comes up which states that the Connection was reset. The page you are trying to view contains POSTDATA. If you resend the data, any action the form carried out (such as a search or online purchase) will be repeated. To resend the data, click OK. Otherwise, click Cancel. The browser which I am using is Firefox and the processing power of my PC is a Pentium II 300Mhz. This problem does not appear when I use a different PC. It takes about 30 seconds for the error message to appear. Moreover, I have also increased the length of timeout in my browser but this problem still persists. Will the problem be gone if I re-install Firefox to my PC? What should I do to remedy this problem? --Siva1979^{Talk to me} 15:14, 12 November 2007 (UTC)

I usually get those timeouts, Wikipedia's fault as far as I know. Is the internet connection the same one as the other PCs? Are you editing a very long page? Have you tried with another browser? -- ReyBrujo 16:01, 12 November 2007 (UTC)

Isn't that something to do with the connection 'keepalive' setting? Have you tried simply clicking "OK"? What happened? SteveBaker 16:11, 12 November 2007 (UTC)

This problem usually appears when I edit long pages. I did not have this problem before. I think that after adding in some add-ons to the Firefox browser, the problem appeared. --Siva1979^{Talk to me} 01:53, 13 November 2007 (UTC)

It is possible. I can never successfully edit List of best-selling video games, which is at 100kb now. Although my edits are saved correctly, the page never reloads, instead I get an error page. In fact, I am already used to it. -- ReyBrujo 04:21, 13 November 2007 (UTC)

Ive gotten those messeges like 700 times in like the last 5 minutes, just click it and it wont do anything. Its mostly my connection though that causes that problom. Esskater11 04:23, 14 November 2007 (UTC)

So, how do I remove these add-ons from my Firefox browser? I re-installed the browser into my PC but the add-ons still remain. What should I do then? --Siva1979^{Talk to me} 14:47, 14 November 2007 (UTC)

To remove firefox add-ons, go to Tools > Add-ons, click on the Add-on you want to remove, and press the Uninstall button. Do this for all the add-ons you want to remove, then close all Firefox windows and restart. You can try narrowing down which one is causing your problem by disabling them one at a time instead (remembering to restart your browser each time), or re-installing one add-on at a time and checking if you are still getting the error.-- Kateshortforbob 22:54, 14 November 2007 (UTC)

I am still getting these errors after doing all these steps. So I guess that the add-ons could not have been the problem. But I did not experience this problem before. This was a very recent problem and I am unable to find the cause or fix this anomaly. --Siva1979^{Talk to me} 03:01, 15 November 2007 (UTC)

Speakers

How do you stop speakers on computers picking up foreign radio stations? Simply south 18:54, 12 November 2007 (UTC)

Um, what? My speakers don't pick up foreign radio stations. Maybe you could describe the problem in a little more detail without presuming that we all know exactly what you are talking about. --24.147.86.187 19:03, 12 November 2007 (UTC)

If you have speakers switched on, they will pick up Indian radio stations or something similar. This seems especially so when you are connected to the internet through a wire. Simply south 19:06, 12 November 2007 (UTC)

Yeah this occasionaly happens, I doubt the stations are very far away though. The only thing you can do is fiddle with the wiring for your speakers, try to use different lengths of wire, or arrange the wires differently, so they stop acting like an antenna. If that doesn't work, get some new speakers. -- Diletante 19:59, 12 November 2007 (UTC)

But they might be with a local transmission station in the area. Simply south 21:10, 12 November 2007 (UTC)

The university-owned computer speakers in my dorm suite last year picked up the frequency of incoming phone calls.. all the speakers in the suite would buzz a full 3-4 seconds and then someone's phone would ring --ffroth 21:24, 12 November 2007 (UTC)

It's caused by interference - the wires to your speakers are probably unshielded and therefore work just like an aerial and pick up signals. Try a slightly more expensive pair of speakers - the ones I have were fairly cheap but I never get any trouble at all - if you spend about £30 (about 60 USD?) you should get a reasonably decent set. The cheaper ones tend to use unshielded cable - if the cable is shielded, it reflects away nuisance external radio waves.GaryReggae 22:25, 12 November 2007 (UTC)

To be fair (and picky), they don't reflect. --Wirbelwindヴィルヴェルヴィント (talk) 22:54, 13 November 2007 (UTC)

Live Messenger down under Vista

Hello, you are my very last try before I shut Vista into the trashbin and return to XP. Live messenger simply doesn't work. I have Vista Home Premium, and for two days after the prchase of the laptop that came with vista installed already, messenger was wrking fine. Then suddenly it gave me the error code 80070005 telling me the service is unavailable, whernever I want to log in. I sent two long nights on google figuring out what to do, registered on several forums like vista and live messenger developers forums but nobody's advice could get messenger work on my comp... Tried reinstalling the messenger, deleting all messenger-related files after uninstalling, downloaded newer and oder versions, checked the firewall setups and did certain cmd tricks people advised me to do, but NOTHING works. Anyone has any idea why my vista just wont let lve messenger work??? I use trillian and miranda to connect but hate them, i want my lil msn back :((( Thanks in advance --Teemeah Gül Bahçesi 18:56, 12 November 2007 (UTC)

MSN sucks :/ Try Pidgin (software) --ffroth 23:06, 14 November 2007 (UTC)

I have miranda and trillian but I don't like them. I want my good oldie msn interfaceeeee --Teemeah Gül Bahçesi 13:41, 16 November 2007 (UTC)

Unix executables and PHP

I'm trying to write a script in PHP that will use the convert binary command-line tool from ImageMagick to convert individual PDF pages into PNG files (rasterize them). I understand that to do this I will also need to have a ghostscript binary at hand as well?

My goal is to execute this with a system command in PHP.

I'm a bit confused as to how the Unix-based filesystem works. In an ideal world I would just be able to grab the few binaries I needed and upload them and execute them (assuming I have the right permissions set, etc.). I can do similar things without too much difficulty on a Windows box, but my server is Unix-based.

I'm trying to code it on my OS X using its built-in Apache server and it just doesn't seem to be trying to execute the "convert" program (I keep getting a "file not found" response from the system command). Here is the kind of command I'm trying to run: ~/Sites/PDFConverter/convert -size 700 "pdf/152511.pdf[2]" "png/152511.pdf.0002.png"

Clearly I'm not thinking about this correctly. Can someone who groks this stuff tell me exactly what I'm going to need to do? I don't want to compile things. I don't want to do anything complicated that I'm not going to be able to do on my server box. I just want to use this command-line program to convert stuff. This is driving me a bit crazy. --24.147.86.187 18:59, 12 November 2007 (UTC)

First, I think you're going to have to use the full path. ~ is interpolated by the shell to mean the home directory of the current user running the shell, which is probably the 'apache' or 'httpd' user, not your user, so it's looking in the wrong place. Also, is the 'convert' program you uploaded built for MacOSX? Are all the libraries for it available? I think you can see if the dynamic libraries for the binaries are available by running on the remote server: 'odump -L /path/to/binary'. -- JSBillings 19:12, 12 November 2007 (UTC)

Maybe you can take a look at here is well, I think you can do what you want directly in PHP through the imagick library. --antilived^{T | C | G} 22:56, 12 November 2007 (UTC)

You really should use the full path and not something that starts with '~'. The '~' means "my home directory" - but the PHP program is being run by the Apache HTTP server and it will probably be running as some other user. Hence '~' means that user's home directory - not yours. (mine runs as the special user account 'www' - so '~' for PHP on my system is /home/www or something. Go with /home/yourname/... instead and all should be well. SteveBaker 01:16, 13 November 2007 (UTC)

Now that I'm on my mac, I see that the command is 'otool -L /path/to/binary', not 'odump'. -- JSBillings 11:30, 13 November 2007 (UTC)

Also bear in mind that ~-expansion is done by the shell. Actually asking the kernel to exec a program called ~/Sites/PDFConverter/convert probably isn't going to get you very far; I don't know whether PHP spawns external processes via the shell, but it's probably worth avoiding using ~ just the same. The Wednesday Island 17:03, 13 November 2007 (UTC)

The Unix 'pwd' command will show the current directory, and might help you figure out in which directory your PHP system command is executing. The PHP phpinfo() function might also help, but you have to know what you're looking for in that. (SEWilco 21:13, 13 November 2007 (UTC))

Concatenating strings in C

I don't get why the string functions in C work like they do. For example, an obvious string concatenation function would work something like this (no error handling, I know, and I haven't even tested it):

char* stringconcat(const char* str1, const char* str2){
    size_t len1 = strlen(str1);
    size_t len2 = strlen(str2);
    char* ret = (char*) malloc(len1 + len2 + 1);

    strcpy(ret, str1);
    strcpy(ret + len1, str2);
    return ret;
}

I can't find such a function, is there one? —Bromskloss 20:46, 12 November 2007 (UTC)

strcat, strncat and similar. -- ReyBrujo 20:56, 12 November 2007 (UTC)

Note that it does not allocate memory like you want, that would be a strdup-like concatenation (which may create memory leaks). -- ReyBrujo 20:59, 12 November 2007 (UTC)

(ec)If you're asking "why doesn't strcat just malloc the data itself, rather relying on the caller to do it" the explanation is partly that the theory of operation of libc (particularly the really fundamental stuff like strings) is not to allocate memory itself, so the caller knows what's getting allocated and takes responsibility themselves. An additional reason is that quite a lot of system level C operates in environments where malloc isn't available (e.g. bootup code, ISRs, and limited-capability embedded devices) or where a special memory-allocation regime applies (e.g. the Linux kernel's kmalloc system, the classic Palm Pilot's persistent storage scheme). -- Finlay McWalter | Talk 21:09, 12 November 2007 (UTC)

Oh, and also malloc() is (generally) not (remotely) time deterministic, which breaks the real-time expectations systems programmers have in some circumstances. The design of libc maximises the number libc calls a programmer can use in places like that. -- Finlay McWalter | Talk 21:22, 12 November 2007 (UTC)

One of the principles of C is that it doesn't do more than you ask. This is important for performance applications. If I want a string concatenation that allocates memory and all I have is one that doesn't, I can still write it in just one line of code:

  char *c = malloc ( strlen(a)+strlen(b)+1 ) ; strcpy ( c, a ) ; strcat ( c, b ) ;

...however, if I want a concatenator that DOESN'T allocate memory and all I have is one that does, then I'm screwed - there is no way to avoid the penalty of 'malloc' - even though I don't need it.

So, given the choice, it's better to provide the simpler functions and let you build on them than to give you complex functions that you have no way to simplify. Of course they could have done both - and arguably that would have been a good thing. You might also argue that you could dispense with strcat at all - since you can write it in C pretty easily - however, that's not really true. Most implementations of strcat use some horribly machine-specific machine code bulk copy thing that you simply can't write as efficiently in C.

C was intended to be only slightly more than machine code and keeping it lean and mean was considered to be A Good Thing. This same reasoning explains why C doesn't do things like array-bounds checking or runtime type checking...if you need it, you can write it, if you don't need it, C doesn't ram it down your throat. This passion for giving you the minimum you need rather than the most explains the phenomenal success of C compared to all of it's long-forgotten rivals. Anyone still writing in RatFor? Algol? CHILL? BCPL? Modula-2? ....No? Didn't think so. A few poor souls are still stuck writing Fortran, Pascal and Cobol - and a few are afflicted with Ada - but mostly for legacy reasons.

These days, C (and it's descendant C++) are pretty much the only compiled languages in common use. Java and Ruby and Python are all interpreted languages - and different considerations apply. SteveBaker 00:35, 13 November 2007 (UTC)

Hebrew on ICQ on a Mac

I have recently purchased a new macbook. I learned how to install hebrew on it, and the hebrew works fine in Word, online, or in other contexts. However, I communicate with friends and family through ICQ and therefore need to be able to type hebrew in ICQ. I have tried to, yet I can not figure out how to get hebrew to work on icq. It just comes up as gibberish and I can not type nor read things in hebrwe. If anyone has any knoweldge as to how to use hebrew, or other languages, on the Mac version of ICQ please let me know. Thanks you

Tammazaf 21:21, 12 November 2007 (UTC)Tammazaf

Try using Pidgin (software) to connect to the ICQ network --ffroth 02:29, 14 November 2007 (UTC)

Does Pigdin have Hebrew support? Guroadrunner 05:16, 15 November 2007 (UTC)

Boot Camp and Neverwinter Nights

IfI have a Mac running Leopard, could I use Boot Camp (and a Windows OS) to play Neverwinter Nights with it's Aurora toolset? —Preceding unsigned comment added by Tuesday42 (talk • contribs) 22:06, 12 November 2007 (UTC)

Yes, you could. Using Boot Camp, Leopard transforms into a full Windows XP or Vista. So you can do anything you can do on a windows. --76.87.166.143 04:10, 14 November 2007 (UTC)

Odd website

I recently came across a website that my friend gave to me [www.internetisseriousbusiness.com] (PLEASE DON"T CLICK ON IT YET). When I went on the site, my firefox browser window shrunk and began to warp around my monitor, all the while a video was playing on the website. When I attempted to close it, repeated pop-ups began to appear. My question is, how does one make a website so that when someone lands on it, it shrinks the user's browser and makes it dance around the screen? Thanks. Acceptable 22:10, 12 November 2007 (UTC)

It's all done with smoke and mirrors...sorry, JavaScript. Javascript can be very powerful and can be used to open and close windows, move them around etc. GaryReggae 22:20, 12 November 2007 (UTC)

Forgot to say, if you go to the View menu then Page Source when you are on the website in question, it will show you the exact HTML code that makes up the website as well as any other code such as JS. GaryReggae 22:21, 12 November 2007 (UTC)

You probably should turn them off anyway, it's in Preferences/Content/(Javascript) Advanced and untick all the checkboxes. For the alert boxes you can use the web developer plugin or the preferences to disable javascript altogether, but otherwise there's nothing much you can do with alert boxes. --antilived^{T | C | G} 22:51, 12 November 2007 (UTC)

That sounds a little overkill to me. I have Javascript enabled and it lets me do all sorts of useful little things most of the time. I can't remember the last time it did something really undesirable, but even then it's not that hard to force-quit a program if I happened to get caught in some sort of infinite javascript loop. I'm not sure, as web users and developers, that disabling things simply because they might be used in a way we don't like is a great philosophy. --24.147.86.187 23:16, 14 November 2007 (UTC)

You can selectively turn off annoying features like maximising the window or move it around, which serve no practical use whatsoever to the user. If I minimise it or move it to a corner I want it to stay there, not jumping around my screen. --antilived^{T | C | G} 23:28, 14 November 2007 (UTC)

Incidentally, you've just been RICKROLLED. :-) --81.157.26.72 (talk) 19:59, 18 November 2007 (UTC)

Window sizes and locations

OK, I've got a problem that has been bugging myself and a colleage of mine for a while. When he opens Internet Explorer (v6 I think), the IE window is always really small and in a corner of the screen. I thought that maximising the window then closing all the IE windows would cause it to start maximised next time but that doesn't work. I then tried dragging the edges of the window so it fills the screen, all well and good but when you exit IE and start it up again, it's back to its usual small self. Why on earth does it do this? Surely it is not that difficult to 'remember' the size of the window when it was closed like a lot of other apps do? GaryReggae 22:29, 12 November 2007 (UTC)

IE is certainly supposed to remember the size of its window, so if it's not doing so it could be some sort of bug. First thing is to make sure you have only one window open, and then size it and close it. A quick google gives a couple of interesting links [1] [2]. Though I'm not sure of the veracity of some of the things they say, their instructions are worth a try. -- DatRoot 23:32, 12 November 2007 (UTC)

Thanks, getting window the right size then holding down CTRL and going to File and Close seems to have cracked it. GaryReggae 12:50, 14 November 2007 (UTC)

HDMI 1.3a cables vs. earlier versions

I often see very expensive HDMI cables touting that they meet the "1.3a" spec. It smells of a marketing ploy much like how some USB cables are marketed for faster USB 2.0 connections. Are 1.3a cables physically different than earlier (1.1, 1.2, 1.3) cables? A digital cable is a digital cable, right? --24.249.108.133 22:49, 12 November 2007 (UTC)

Well there is a bit of a debate that rages on with digital-cabling as to whether improved quality cabling improves performance. One side suggests that digital doesn't change from the cheapest to most expensive as they use 0/1s as 'what you put in you get out' and anything less would cause failure...(http://www.statusq.org/archives/2007/01/24/1287/) for example suggest there is no difference to be found as does (http://gizmodo.com/gadgets/hdmi-cable-battlemodo/the-truth-about-monster-cable-+-grand-finale-part-iii-282725.php) but this article (http://www.hdtvmagazine.com/articles/2006/08/hdmi_part_8_-_c.php) covers the 1.3a bit at the end by the looks of it. It's all far too complex to get my head around but i suspect that it isn't quite clear cut either way. Presumably a system that receives a bad-signal will 'return' requesting another copy, if this is the case then more reliable cabling might make less 'returns' making the system work less hard, thus work better. ny156uk 00:05, 13 November 2007 (UTC)

Sorry just found this too (http://www1.electronichouse.com/info/specials/hdmi_basics.html) that backs up the 'it makes no difference' idea. Also just realised your question is more about is 1.3a different - oh well i've written the reply now. ny156uk 00:09, 13 November 2007 (UTC)

This is similar to the debate between optical (S/PDIF) and CoAx for audio. I've tried both and there is absolutely no difference and I have a pretty decent midrange system (Wharfedale). Similarly, I've compared HDMI to ordinary RCA (red/white/yellow connectors) and noticed no difference in the DVD playback quality. The only time I could ever hear an audio difference was when I plugged in my laptop to my HiFi and played mp3's. The optical source from the HiFi cd player, whether it was mp3 or cd, was a lot better. Sandman30s 21:19, 13 November 2007 (UTC)

using /dev/random to generate typeable characters

How can I get a random stream of typeable characters from /dev/random. things like cat /dev/random | uuencode only gives me a subset of the typeable characters. I am using this for password generation so I don't want things like linefeeds, accented characters, etc. -- Diletante 23:00, 12 November 2007 (UTC)

ok so I found out that a good solution is to pipe to tr like cat /dev/random | tr -cd "a-zA-Z0-9-_~\!@#$%^&*()=\$\?" . It seems to work like I want.--Diletante 00:17, 13 November 2007 (UTC)

I sure hope you aren't even considering inflicting random passwords on your poor users?! It's well known that random passwords are very insecure because nobody can remember them - so everyone writes them down and sticks them onto a post-it note in their desk drawer! SteveBaker 00:19, 13 November 2007 (UTC)

Good point, this is for my use only. -- Diletante 00:50, 13 November 2007 (UTC)

I still generate random passwords for user's initial password, and tell them to change it immediately. -- JSBillings 11:25, 13 November 2007 (UTC)