Wikipedia:Reference desk/Archives/Computing/2009 September 23

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.

September 23

How do you quickly familiarize yourself with code written by someone else?

Say you have the source files for a software application but there are no design documents, and you want to modify the software. What steps do you take to help yourself understand the organization of the code? Are there tools that you'd recommend for facilitating the analysis? —Preceding unsigned comment added by 173.49.10.73 (talk)

Maybe a look at Michael Feathers' book Working Effectively with Legacy Code (ISBN 978-0131177055) might help. —Tobias Bergemann (talk) 10:14, 23 September 2009 (UTC)

I find adding print statements is quite valuable: "Oh, so this chunk of code is hit whenever the value of variable X is in the Y-Z range, I see...". StuRat (talk) 13:27, 23 September 2009 (UTC)

I go through and change the indentation and use of { and } to my own preference, one line at a time. It forces me to scan all of the code. By the time I'm done, I have a pretty good idea what the code does. -- kainaw™ 17:12, 23 September 2009 (UTC)

Charging Li-on batteries

Will the quality or lifespan of lithium ion batteries degrade if left charging for long periods of time? Assuming the battery has the battery protection circuit. -- penubag  (talk) 06:19, 23 September 2009 (UTC)

If the "battery protection circuit" prevents it from charging when already fully charged, then I don't see how being plugged in all the time would have any effect at all. This Q is very similar to the "Charging my laptop" Q a few down from this one, so also look for replies there. StuRat (talk) 12:52, 23 September 2009 (UTC)

Thanks for the heads up -- penubag  (talk) 01:58, 24 September 2009 (UTC)

HELP with my rogers phone to a new one im supposed to get!!

hello,

I currently have the KRZR from ROGERS. and it has a spot for my SIM CARD. i was looking to buy an HTC TOUCH DIAMOND.. unlocked.. and i thought i would be able to pu tmy rogers Sim card in and it would work? whats am i supposed to do!?

HELP PLEASE! (:

MathML support in Mozilla Firefox

What versions of Mozilla Firefox support MathML? I used to use MathML with some older version, but my current version (3) does not render MathML. Does 3.5 support MathML? Thanks! --Masatran (talk) 11:17, 23 September 2009 (UTC)

Firefox 3.0.14 (linux x86 ubuntu) passes the MathML test at https://www.eyeasme.com/Joe/MathML/MathML_browser_test pretty well (with a few missing-font issues). Maybe you're missing fonts too? -- Finlay McWalter • Talk 11:25, 23 September 2009 (UTC)

EasyCap 2.0 USB problem

Hi everyone I wondered if somebody could help me with a problem with a friend's capture card. My friend's computer OS is Windows XP and the capture card is called EasyCap 2.0. The default setting to record is NTSC_M but the one my friend wants is PAL_M. Every time he goes to change it, it just goes back to the default NTSC_M setting. This means because his TV and Wii is PAL the capture card can only record in black and white...Is there any way to change the setting permenately to PAL_M rather than NTSC_M so he can record in colour? Will continue to monitor this section for any other answers...Thanks Chevymontecarlo (talk) 11:36, 23 September 2009 (UTC)

Charging my laptop

Is it best to keep my laptop plugged in 24/7 or to periodically charge it and allow it to discharge ? I am a college student, so mostly my laptop is on all the time. I am asking this question from the point of view of the battery... I have heard that its not good for cell phones to be kept on charge always, but I'm not sure if the same applies to the laptop battery,....Rkr1991 ^{(Wanna chat?)} 12:20, 23 September 2009 (UTC)

I think the most serious problem is devices that continue to try to charge the battery even when it's already fully charged. This results in it getting hot, which in turn shortens it's life. I'd expect an expensive device like a laptop, which already has the ability to monitor the charge level of the battery, to stop charging when the battery is fully charged. You can test this for yourself by feeling if the battery is still hot after it's been fully charged for hours, but not turned on. StuRat (talk) 12:46, 23 September 2009 (UTC)

I have a thermal monotor built in. It is around 50 (in degrees centigrade) when idle and around 70 during use... So I think that means it's pretty safe to keep my laptop plugged in, doesn't it...? Rkr1991 ^{(Wanna chat?)} 12:50, 23 September 2009 (UTC)

That's within the normal range for a processor. --98.217.14.211 (talk) 13:03, 23 September 2009 (UTC)

Yes, but we need to know the temp of the battery specifically, not the processor. What's the battery temp when charging and hours after fully charged but still plugged in ? StuRat (talk) 13:20, 23 September 2009 (UTC)

As far as I can tell, it doesn't matter much. The battery will, at some point, start losing its effectiveness and, if you have it long enough, may need to be replaced. But this happens no matter what you do, as far as I can tell. --98.217.14.211 (talk) 13:03, 23 September 2009 (UTC)

I use a laptop as my weather station and I leave it permanently on and plugged in - 24x7. My current one has been doing this for about a year. The battery still works fine (when there's a power cut). --Phil Holmes (talk) 13:33, 23 September 2009 (UTC)

Well, but there's a lot of range for "works fine," though. There are usually utilities you can get that tell you the health of your battery. Mine is 85% at the moment (that is, it can now only hold 85% of its original charge) -- not great, but not worth replacing at this point. I've had them dip pretty low after some years. If you don't use it on battery very much, you probably won't notice it, but if you do use it a lot, you start to notice it holding less of a charge after awhile. --98.217.14.211 (talk) 15:26, 23 September 2009 (UTC)

I remember reading an Apple document not too long ago that advises use of the battery at least once a month and to leave the battery at half charge if it is going to sit for multiple months. I assume this is the case for all Li-ion/Li-poly powered laptops. If you take your laptop off your desk and off the power adapter every few weeks, your battery should be good (mine is plugged in the majority of the time and it's stil at 95% health after a year and a half) Caltsar (talk) 16:05, 23 September 2009 (UTC)

As far as I'm aware any worry about 'leaving the battery charging' is unfounded - laptops use a lithium type battery (which are not far off fireworks as far as energy density goes) - thus they have built in electronics which prevents things such as overcharging etc. A full battery can be easily detected since the resistance (at a given charging voltage) rises rapidly once it is full. However I have read that it's a good idea to 'power cycle' these batterys (some instructions mention it) - however this is a different topic. The article Lithium ion battery also mentions built in thermal shutoff in case of overcharging.83.100.251.196 (talk) 17:34, 23 September 2009 (UTC)

Also note that if the device is permanently plugged in it's not using the battery, but drawing power from the power supply.83.100.251.196 (talk) 17:37, 23 September 2009 (UTC)

Are you positive that this is the case for every laptop? There may be laptops out there which always take current from the battery, even when it's plugged in, yes? Comet Tuttle (talk) 20:49, 23 September 2009 (UTC)

I doubt it - think electrical safety requirements, also see [1] - manufacturers of cells require buyers to use safety electronics, otherwise no-sell - (they don't need arseholes making unsafe batteries from cells - causing accidents - and thus making the cells unpopular etc) - Li batteries require special charge control to limit the charging current (so to prevent overheating and BOOM!).

Also see [2] section 'safety circuits' 83.100.251.196 (talk) 21:33, 23 September 2009 (UTC)

Oh you meant it draws power from the battery even when plugged into the mains - ok -why would it do that?83.100.251.196 (talk) 21:36, 23 September 2009 (UTC)

Because it's cheaper, since no switch is required to go from one power source to another. I have a portable DVD player and rechargeable walky-talkies like that, they won't run without the battery in place. And even with the battery in place, you still have to charge it up a bit before it will run. It's a sucky design, but since the consumer has no way of knowing if it works this way, prior to purchase, they don't know to avoid such devices. And, after purchase, unless they try to run it without the battery in place, they may not notice that it works that way until after the return period has ended. StuRat (talk) 14:29, 24 September 2009 (UTC)

Yes, but you can't draw power from a battery and charge it at the same time.83.100.251.196 (talk) 18:02, 24 September 2009 (UTC)

This link contains some relevent info about Li batteries [3].83.100.251.196 (talk) 17:40, 23 September 2009 (UTC)

Having left my laptop plugged in almost permanently for over two years, I note that its usable time has reduced from three hours to one hour. I discharge and recharge it occasionally (once or twice a month). Is this reduction in usable time just normal deterioration, or would it have lasted longer if I had not left it on permanent charge? The advice of the manufacturers is not to leave it permanenly plugged in, but I usually ignore such advice. Has anyone done any research on what difference it makes? Dbfirs 12:51, 24 September 2009 (UTC)

Is this code unbreakable?

Hi all,

In order to encrypt various bank details and passwords on my computer, I created a program that uses a simple Polyalphabetic cipher. I then generated a key that was longer than my message (using all existing ascii characters) and encrypted the message once.

Am I right in thinking that, if the key were truly random, this would be unbreakable? In the section here it says that it's "difficult to break" if the key is less than 27 times the length of the plain text, but it doesn't say whether having the key being longer than the plain text makes it unbreakable.

In practice, my key was not random, but produced using Java's Random algorithm, which takes the time as a seed (I believe). How much less secure does this make the code?

Thanks! — Sam 63.138.152.155 (talk) 16:06, 23 September 2009 (UTC)

If your key is as large (or larger) than your plaintext, and you never reuse the key material, then you have a one time pad. A OTP is, formally, provably secure; but in practice using a OTP properly is very difficult (as there's so much key material to store and manipulate) that it's only used for very rare things like diplomatic wires. java.util.Random is not a cryptographically secure random number generator. Your scheme founders on the practicalities alone - rather than keeping safe an N-character block of blank info, you now have to keep safe an N-character block of key material, which is essentially the same problem. If you're genuinely intending to use this to store real secret data, use GNU Privacy Guard, Password Safe, or KeePass rather than brewing something yourself. -- Finlay McWalter • Talk 16:14, 23 September 2009 (UTC)

As to the specifics of the random number generator, see Cryptographically secure pseudorandom number generator. A regular pseudorandom number generator is insecure because its period (how long before it repeats itself) is too short; and in practice because it may have unpleasant algebraic properties that allow an attacker a shortcut. -- Finlay McWalter • Talk 16:21, 23 September 2009 (UTC)

Answering the last question... If I know the algorithm you used to make your key, and I know that it was created using Java's random function, and I know it used the time as a seed, and I know approximately when you created the key... I can create keys using every time from before I know you made the key until after I know you made the key. One of those keys will be your key. I try each key on the encoded text until one works. This gets into automated brute force techniques. I obviously won't actually read the output of each key. I will send that into a pattern-matching filter. If it matches the pattern of what I expect the plain text to be, then it is saved off for me to read. -- kainaw™ 17:09, 23 September 2009 (UTC)

Pedantry: Generally speaking, the answer to your title question is "no". If the core technique is sound, an attacker is likely to simply go around it. What if your program swapped the plaintext out to disk? What if there's a bug in your program? What if the NSA is using TEMPEST techniques on you? What if your spouse is a mole? What if Dick Cheney tortures your loved ones until you give up the key? "Unbreakable" is a very high bar. --Sean 20:03, 23 September 2009 (UTC)

This is kind of an interesting question. Lets put aside rubber hose techniques (that's not what he's asking). Just because your cipher's polyalphabetic doesn't mean it's analogous to a OTP. I'm going to assume you actually have access to true random numbers. The better question is why, if you have a key longer than your plaintext, do you need an encryption method other than OTP. OTP is provably secure. If you can keep the key secret/remembered why can't instead keep the information secret? If you somehow can remember the key better than the original data then I suspect the original key isn't random (or you got "lucky" and your key is probably susceptible to some other attacks) . Shadowjams (talk) 09:23, 25 September 2009 (UTC)

Commodore 64 to a modern TV?

My team mate at work saw me wearing a Commodore 64 themed T-shirt and said he had a fully working Commodore 64 for sale. Now this made me intrigued about whether it's possible to connect a Commodore 64 to a modern flatscreen TV. The Commodore 64 uses an RF modulator for video output. Back in the 1980s, it was standard for all TVs to accept a signal from an RF modulator (it was often the only input available), but is it used on modern TVs any more? JIP | Talk 16:20, 23 September 2009 (UTC)

Sure, my year-old Sony flatscreen has an RF-input, which can be used to watch analog TV (so there's no reason a C=64 or whatever wouldn't work). Once digital switchover is complete, there's much less reason for TVs to continue to feature that, so it'll die out over the coming years. -- Finlay McWalter • Talk 16:23, 23 September 2009 (UTC)

Great! Now I'll just have to buy my team mate's Commodore 64 (which I think will cost a negligible sum) and a flatscreen TV (which will probably cost about 200-300 €). Or then bring it around to my father's place and use his really fancy wall-mounted flatscreen TV. JIP | Talk 16:24, 23 September 2009 (UTC)

The Commodore 64 has both an RF port and an AV port. The AV port carries mono audio, composite video, and luma/chroma video. If you can find or make an appropriate cable, it's possible to connect the Commodore 64 to your TV's composite video or S-video input.

According to Commodore 64 - IO ports some early Commodore 64s have 5-pin port with only composite video. To make sure, just check your Commodore 64's AV port (next to the RF port). If it has 8 holes it's the newer type. If it has 5 holes it's the older type.

Be careful when shopping though. Original Commodore 64 AV cables designed for Commodore monitors use a yellow cable for luma (a grayscale picture), a red cable for chroma (color information), and a white for mono audio. If you plug these cables into a stereo TV with yellow, white, and red ports, you'll get a grayscale picture, correct audio from the left speaker, and buzzing audio from the right speaker.

Also, an AV cable for a Sega Genesis uses the same connector, but different cable assignments. We had one and when we connected it to the Commodore 64, I remember we could only get a grayscale picture with it, but I forget if we got sound or not. --Bavi H (talk) 03:31, 24 September 2009 (UTC)

You can get perfect C=64 emulators online for free right now (such as VICE); they do everything the physical one will do, and (because disk and tape files load nearly instantaneously) you don't have to put up with those twenty-minute "fastload" flickery-border loading times. The real thing probably isn't going to be as fun as you remember ;( -- Finlay McWalter • Talk 16:29, 23 September 2009 (UTC)

I already have VICE installed on my Fedora Linux system, and it works perfectly. But I would like to be able to transfer the about 500 original Commodore 64 disks I have to the system. I went out and bought a 1541-to-serial adapter cable online, but only after I had completed my order did I realise my system doesn't have old-style serial ports. Direct 1541-to-USB adapter cables are still in hobbyist prototype stages, but are there serial-to-USB adapter cables available for sale? JIP | Talk 17:21, 23 September 2009 (UTC)

Sure, just search for "usb serial port adapter" at your favourite retailer; they're cheap and plentiful. -- Finlay McWalter • Talk 17:24, 23 September 2009 (UTC)

I'm not so sure about that. All hits for "USB serial port adapter" I could find are for 9-pin RS-232 serial ports. The cable I ordered is for an even older variety of serial ports, with something around twenty pins. Also I'm worried about cable gender issues, I might end up trying to smash two connectors with the same gender together. JIP | Talk 18:23, 23 September 2009 (UTC)

That larger format is a DB-25; the mapping of pins on the DB-25 and DE-9 is given at RS-232#Pinouts. You either need a 9-25 adapter or you can build one - in practice you almost always only need the GND, TxD, and RxD lines - the rest are meaningful only for modems. If you're doing that yourself you may have to wire things as a Null modem (which essentially means one side's TxD is the other side's RxD and vice versa). -- Finlay McWalter • Talk 18:31, 23 September 2009 (UTC)

I would very much like to avoid having to build a single cable myself. The last time I soldered something was back in primary school, and I did a pretty bad job at that too. JIP | Talk 19:31, 23 September 2009 (UTC)

Are you sure your 1541 transfer cable is for a PC serial port? The 1541 transfer cables I'm familiar with are for a PC parallel port (printer port). For example, see sta.c64.org/cables.html. If I'm interpreting it correctly, all of the disk transfer cables listed there connect to the PC parallel port (printer port). For the connection on the other end, most of the cables listed there connect to the Commodore 64 or Commodore disk drive serial port (the standard Commodore disk and printer bus), and so they're called serial cables. However, if you modify your Commodore disk drive, you can install a special parallel port on it and transfer files faster. The cables that use this special parallel port are called parallel cables on that page.

Also, I made some comments about AV cables above. --Bavi H (talk) 03:31, 24 September 2009 (UTC)

It's an XM1541 cable, and yes, it actually seems to be for a parallel port. There also seem to be parallel-to-USB adapter cables available. I'm still worried about cable gender issues though. JIP | Talk 04:34, 24 September 2009 (UTC)

After some Googling, it turns out that the real problem is in the parallel-to-USB adapter, but of a completely different sort. Modern parallel-to-USB adapter cables are stripped-down versions that only use those pins needed by printers. The XM1541, on the other hand, makes full use of the parallel cable, and thus expects to find a complete parallel port on the other end. This means I need a USB-compatible parallel port replicator, which seems to be extremely hard to find, and even I do find one, it will cost me over 60 €. Damn, it feels like the whole world is trying to stop me from accessing my old Commodore 64 software. JIP | Talk 17:31, 24 September 2009 (UTC)

If you bought all that software, there should be nothing stopping you downloading images of them. It will be far easier than trying to transfer them all, assuming you can still read more than 50% of those old mouldy disks... Sandman30s (talk) 12:47, 28 September 2009 (UTC)

VPN, certificates, and CAs

I'm setting up a Netgear VPN router for the first time and the Netgear manual talks about getting a certificate from a CA, and uploading it to the router. Do I really need to do all of this? If I do need to do it, who's the cheapest of the reliable CAs? What company do most people use? Comet Tuttle (talk) 17:56, 23 September 2009 (UTC)

The cheapest CA would be the one you would set up yourself. In the netgear website i were able to find only one VPN router and it seemed to be rather high end device (8 gbit ethernet ports, etc), and it uses IPsec. I have never dealt with IPsec VPN's, but SSL VPN usually can either be set to use preshared key or no encryption at all (both are considered less safe than use of CA and certificates). Maybe there is available something similar. -Yyy (talk) 10:45, 24 September 2009 (UTC)

open modem

i have a Huawei 3g Broadband Dongle Modem..it's locked onto using one 1 service subcriber but on googling i found out you can open it to use whichever network available..There's a place that you send your imei number and they send you the unlock code..i thot i could save and donate the amount ..i downloaded the software from http://wareseeker.com/download/huawei-modem-unlock-code-tool-v1.1.rar/7cddfaaf4..but it's not installing..Anyway i need help if anybody knows how i can unlock my modem..the imei is XXXXXXXXXXXXX...so help.

I've removed your IMEI from your post just as a precaution. Last thing you want is someone cloning your sim card and running up a huge bill at your expense. I am using a Huawei E220 USB modem (in fact I'm connected from it now). It came with proprietary dialer software, however I can't confirm if it was network locked. Soon after I got it I simply downloaded the latest software and firmware update (two separate entities as I discovered) from Huawei's website, and now instead of the proprietary dialer popping up, the standard Huawei dialer comes up instead, and instead of connecting at 3.6Mbps the connection reports that it is connected at 7.2Mbps. I'm not sure if flashing your firmware with standard Huawei firmware will get rid of a network lock, but it's worth a try. Zunaid 22:25, 23 September 2009 (UTC)

SIM card specific data is IMSI, IMEI is specific to device (this case - modem), so it is not possible to clone SIM card by knowing that. (knowning IMSI also would not be enough to clone SIM). -Yyy (talk) 09:56, 24 September 2009 (UTC)

On the economics of free software

How do Free Software activists like Richard Stallman (as opposed to Open Source adherents) who would that there were no such thing as proprietary software expect full time software developers to make money to support themselves? You always hear "free speech, not free beer," but if beer were distributed with the GPL and could be replicated as easily as software is, beer would realistically have zero price almost instantaneously too.71.161.40.225 (talk) 21:43, 23 September 2009 (UTC)

From the Free software article "Free software business models are usually based on adding value such as applications, support, training, customization, integration, or certification." To extend your analogy, beer is not distributed for free, but the instructions for making beer are. Assuming you have minimal equipment and relatively low cost ingredients, you can make beer. In much the same way, you can make software, but even when the instructions (source code) come for free, there are many reasons why you might choose to buy software & service rather than brew it yourself.

The model appears to work very well for GNU/Linux; per Linux#Community, there are a great many contributors to the Linux kernel and the remainder of the operating system ... I'd hazard a guess that at least as many if not more are engaged in this pursuit as are engaged in developing Windows operating systems. So there is empirical evidence that the free software market does, remarkably, financially support a great many developers.

I grant you it is all seemingly counter-intuitive, but that is merely because we are so wedded to and familiar with proprietary models as to have become conditioned to think only within that paradigm. And that's where hairy monsters like Stallman come in; presumably one of the reasons that he was awarded a MacArthur Fellowship.

Finally, and even more counter-intuitive, is that one can create software and make it available under a free software license and sell it - check out this Q&A with Stallman, who says "So I see nothing wrong with charging money for a copy of a program. However, I do see something wrong with denying the user of a program the essential freedoms, after he’s gotten his copy, whether he’s gotten it by paying for it, gratis, or however he got it." --Tagishsimon (talk) 23:08, 23 September 2009 (UTC)

I'm not sure if they do make any money. I read that 98% of MySQL customers are DIY.^[4] The MySQL division only earns about $50 million a year from support services. SUN paid about $800 million for MySQL in 2008, even though it's a free product. I'd like to see some data relating to other open-source products, but I think it's damning that so few companies open source their software today.--Drknkn (talk) 23:46, 23 September 2009 (UTC)

Damning? What sort of talk is that in the context of a discussion of the economics of FOSS? I really do fail to grasp your point, and just do not even begin to comprehend your tone. I recognise we are on ground that for many appears to be ideological, but the ability of the FOSS market to fund itself is surely the relevant empirical evidence here. That much software is proprietary demonstrates that that model has legs, too. But the one having legs does not delimb the other. Meanwhile, if you want to bandy figures, let's consider Redhat's $652.57 million revenue in 2009 (according to our article). They're but one of a number of suppliers in the Linux field, and it seems to work for them. --Tagishsimon (talk) 00:03, 24 September 2009 (UTC)

As a manager (not an open-source advocate), your first obligation is to your shareholders. (They're the owners of the company.) You're not running a charity. You're playing with other people's money. That's what I meant by "damning." Also, what about SUN and Novell, which are both losing money even amid R&D cuts? SUN has been losing money since 2002. Around the same time, they gave away much of their intellectual property. Novell got out of the Netware and WordPerfect business and into the Linux business, which hasn't worked out very well for them, either. Compare their performance to that of Oracle, Microsoft, and Apple. All of those companies have been doing great with closed-sourced code. RedHat charges for RHEL, and they get away with it because Fedora releases are clearly beta-quality. SUN, on the other hand, charges for StarOffice, even though OpenOffice is just as good. The business model for a while now has been to give away crippleware and charge for software. You guard your source code, because if it's open sourced, then there's no way out. Once it's on the Internet, it stays there. That's why many Oracle shareholders are a bit upset about the acquisition of SUN. I don't blame them.--Drknkn (talk) 01:16, 24 September 2009 (UTC)

I personally am not convinced that the economics works out, but comparing them to Oracle, Microsoft, and Apple—maybe the top three most profitable companies in the business, all of which go WAAY beyond "proprietary" to the level of "lock-in", does not make much sense. Looking at the fortunes of one company does not tell you about an entire business model. There are a lot of reasons that people lose money or gain it; extrapolating from the numbers alone doesn't tell you anything at all. --98.217.14.211 (talk) 14:28, 24 September 2009 (UTC)

Apple is an interesting case since their kernel is mostly open source and partially based on open source components (MS also uses open source components in Windows but not to such an extent AFAIK) but their products are generally heavily restricted Nil Einne (talk) 15:13, 24 September 2009 (UTC)

I think it makes some sense to compare the leading companies in a sector to the leading companies in another sector. It doesn't tell you a lot, but it tells you something. Comet Tuttle (talk) 17:41, 24 September 2009 (UTC)

Skipping out on volunteerism as your primary source of labor (which can be significant, as Wikipedia can attest), it comes down to selling subsidiary products and services. It's certainly not the most profitable model, and would not appeal to big software combines unless there were other advantages to being "free" (like market dominance). Stallman does not frame his argument in economic terms primarily—it is an ideological, rights-based argument first. Any attempts to find ways to make it profitable are secondary. (As compared to Open Source, which is usually presented as a quality-based argument first, rights-based second.) --98.217.14.211 (talk) 23:51, 23 September 2009 (UTC)

Agreed. It comes down to if your primary goal is to make money, with providing quality software to as many people as possible only a secondary consideration, or vice-versa. Although in the case of a company which only provides software as a ancillary service to their real product, both goals can match up nicely. For example, Netflix' primary business is distributing DVDs through the mail, but they also have a rather extensive web site and database system set up to support this. As it happens, some of their software seems to be rather poor. Now, they could hire lots of highly competent programmers to fix it, or they could release all their software to the public and let them fix it. They seem to be trending more to the latter, which could save them money and provide better software in the long run. Now for an example of a business that doesn't do this (AFAIK). When you buy a car it has software to control how the radio, heat and A/C, and other settings work. From my experience there's no easy way to reprogram this software. Thus we get window defrost settings that automatically turn themselves off every 5 minutes, at great annoyance to the customer, who has no way to fix it. If we could just go online and grab a fix for that, put it on a USB flash drive and plug it into our car, then the customers would be much happier and more likely to buy that car again. However, I wouldn't allow user access to any safety related programming, like the brakes or air-bag deployment, that should be left to the "experts".

Now let's talk about companies that are primarily software companies. There I think a compromise is in order. I like the model where the core software is proprietary, but they also give away lots of APIs, GUIs, etc., so that end users can thoroughly customize their software. This hopefully allows the company to make a profit and also allows a large base of application programmers to make the software fit their business and personal needs. StuRat (talk) 19:46, 24 September 2009 (UTC)

Removing "Personal Antivirus" trojan

A friend has gotten themselves infected with the "Personal Antivirus" trojan. I've looked online at the various fixes and noted with alarm that they're often quite different from one another. For example, this site has a list of DLLs that need to be unregistered while this one does not. Even the lists of processes to end don't match up. If I've read things correctly, it looks like there are no free AV fixes for this infection out there, so I'm going to do it manually, but don't know which site has the straight dope on fixing the problem. Is there a site that the pros use to get their removal tips from? Matt Deres (talk) 22:10, 23 September 2009 (UTC)

I've dealt with PAV on several occasions. I simply boot into safe mode by pressing F8 during start-up. Then, I delete the Personal Anti-virus folder inside the Program Files folder. Then, I go into the C:\WINDOWS and C:\WINDOWS\system32 folders, arrange the files by creation time, and delete the newest files. Viruses can mutate, so the file names may change. That's why I simply remove files created since the infection. Then, I search the registry by running regedit and pressing CTRL + F, then typing in the names of the files I deleted. I delete any keys with those names in them. Then, I disable the PAV add-ons inside Internet Explorer. Finally, I delete all temporary files on the computer -- including the folder used by System Restore (C:\System Volume Information, C:\WINDOWS\temp, etc.). As for those instructions, services.exe and winlogon.exe are critical system processes, and I don't know if Windows will function without them or even if you will be able to stop them. In any case, viruses are programs, and programs can crash if they're missing any of their modules. So, if you just get rid of some of the files created by PAV, then I think their computer will work much better.--Drknkn (talk) 22:24, 23 September 2009 (UTC)

Also, another thing to keep in mind is that once a system is compromised, if you really care about privacy, you may want to just do a clean install of the operating system. It takes way too much work to be fully certain that you have gotten rid of some bug. Brusegadi (talk) 06:59, 24 September 2009 (UTC)

Thanks for the advice, guys. Matt Deres (talk) 21:40, 25 September 2009 (UTC)