CoffeeFiend

You're the one who purchased twice as much RAM than your existing x86 OS can use. The x64 "flavor" of Vista and Win7 would use all 6GB just fine (audio would pretty much work out of the box too). As for the SATA hard drive not being seen, it's most likely because of a BIOS option like AHCI mode, which XP won't see out of the box (requires drivers) unlike newer versions of Windows. Or perhaps it just failed to auto-mount with a drive letter (check in in disk management MMC snapin i.e. diskmgmt.msc) Migration of OS to another drive should be easy if it's set to IDE emulation in the BIOS. Not that I would personally run XP on a modern box like that. As for Symantec 360, you couldn't pay me enough to install that junk on my system.

Well, I'd sure want Vista's many new features, but Win7 has plenty of them too. There's plenty of fairly major features like the new taskbar (with jump lists and all), aero snap (and new keyboard shortcuts), homegroup, libraries, multitouch (monitors for it are coming), XP compat mode, plenty under-the-hood changes like support for SSD TRIM, etc. And there's plenty of very nice minor ones too, like desktop slideshow, gadgets can be anywhere, improved start menu, the uncluttered systray area, out-of-the-box DirectX 11/IE8/WMP12, the improved media center, new IIS, improved firewall, improved apps like calc/paint and so on. It all adds up pretty quickly. It's a fairly decent upgrade from Vista, mind you I never had any problems with Vista either. But hey, you can try to downplay it all, and say none of this is worthy, just like loads of people said about Vista vs XP. There's much more to it than just that. Win 7 isn't just some GUI changes. There's a unbelievable amount of new features over XP, even in Vista. I'd never want to go back! It's a modern 64 bit platform with a bright future too.

@Andromeda43: congrats on managing to pick the two very worst tools ever to make web pages. Words' output is 100% pure garbage. Frontpage has a long history of creating crappy code, markup that doesn't validate, stuff that only properly renders in IE and so on. Thankfully it's no longer sold (it was EOL'ed a few years ago), and it was quite overpriced for what it is. If you really insist on WYSIWYG tools (HTML is easy to learn, even for mere mortals), then the 2 main options are Adobe Dreamweaver CS4 and MS Expression Web 3. They're modern tools that don't actually generate pure garbage.

No. It's going to get wiped regardless. As for how to recover, check Intel's own RAID 10 Volume Recovery guide.

That's not really surprising. Without VT-x or AMD-V (mind you, AMD doesn't turn stuff like this off unlike Intel) you can still run 32 bit OS'es under vmware and various others. Mostly everything but Win7's XP compat mode. You could do something similar with the vmware player and various other apps. The main difference here, is that when you use Win7's XP compat mode, it already comes with the XP license to run inside the VM. If you use something else, then you need a separate XP license just for that. Where it really sucks, and where there are no other options besides buying a new non-crippled, expensive-ish CPU (which sucks a lot if you still have a socket 775 mobo i.e. most people still), is when you want to run 64 bit OS'es in VMs Running the x86 version of Win2008 R2 seems kind of silly. Edit: Cool. I had missed that. However, I definitely cannot use the de-de version

Sadly, I have no answer as for which programming language as I don't really mess around WinPE (I have no need to). However, keep in mind most people who want to do something along those lines will opt for other very well known solutions like acronis trueimage or norton ghost (or countless others), unattended installs, windows 7's built in backups, the backups from windows home server and so on (many options which are well known, well supported, mature, fully featured, decently priced and all). And a large percentage of folks already have such a recovery partition on their boxes in the first place (placed there by the OEM, be it Dell, HP, Lenovo or whatever) This leaves you with a very small potential market for a commercial product (especially since we have to reinstall far less often these days), especially if they have to be able to provide their own WinPE source (download the latest WAIK first?) and then run your program from elsewhere manually and all that stuff (doesn't sound all that simple compared to other options) since you can't legally redist any of those files (WinPE) yourself. Just saying, I'm not sure if I'd spend too much energy on this particular project. It's not that easy making decent money as a micro ISV, and here, you'd also have to provide support for it and everything else (yes, even folks who will pirate it will dare to ask!) which most people don't really like doing. Then don't forget all the sales (transaction processing fees, taxes and whatever -- dealing with the chargebacks too) and marketing stuff (e.g. adwords or what not -- if nobody hears about it, you'll have zero customers) and so on. Don't forget the hosting fees (including domain name) for your sales/support site either (nor the time to build it, time to answer emails, time to police support forums for spam and such). I just don't think you'd be making a lot of money per hour worked on the project unfortunately -- call me a pessimist if you want! Starting out as a micro ISV is hard, but starting in a very crowded market is much harder. If I was in it for the money, I'd be picking another niche (without the already overcrowded market)

It is, for both (and there is no such thing in XP indeed). So there's exactly zero change. We've all seen such "I found a huge increase in performance on my system" claims hundreds of times before as well, like the super-duper "enable superfetch in XP" tweaks and the like (nevermind the code that would use that key doesn't exist in XP). There's always a LOT of folks who'd swear it made a gigantic, day-and-night difference when there is none whatsoever. And here, even if it was disabled by default (which it's not), it wouldn't help "global" system performance anyways, just network speed, and only under certain conditions (depending on link speed and latency mainly). There are even folks who had network performance problems caused by this and got a speed boost by disabling it. And like pretty much everyone said already, optimizers are useless junk.

OpenDNS has some advantages, like it lets you block categories of websites (adult content, gambling sites, etc and blacklist some directly) and some nasties, it corrects some common typos (instead of redirecting to those malware and ad-ridden parked pages we all love), and their network latency is better than a lot of ISP's. Google DNS... Might be faster than some ISPs latency-wise. But personally, Google already gets most of my search results (and which ones I've clicked on too), all of my personal emails, locations I'm traveling to (maps) and tons more personal infos. Call me paranoid, but I'm just not handing them the complete list of every single website we visit along with that (with time stamps and all no less). Yet another Google product that thoroughly fails to impress, but manages to get that much more extra data on us nicely...

Hmm, no. Everything defaults to WPA or WPA2 these days. WEP is really old and very insecure. Misusing a range of public addresses will only mess things up quite badly. Not that it would offer any real protection as all your traffic would be broadcasted unencrypted. Just use WPA or WPA2 and a decent password and you're set. Problem solved. Nope. Linksys' stock firmwares aren't exactly loaded with advanced options... There is AP isolation which will isolate WLAN clients from each other, and that's about it. You won't find advanced features like these on stock firmwares of consumer routers, even the expensive ones (my WRT160N is surprisingly awful). Firmwares like DD-WRT and Tomato have a LOT more advanced options, but their hardware support is limited.

Ah, that makes it a lot easier already. For only 30 users and such basic stuff, I'd be going with a fairly "basic" server (not low end or low quality -- you still want it to last) running Windows SBS 2008 Premium (R2 isn't out yet). That will very easily run everything you were asking for (including user profiles, file shares and so on), and it even includes Exchange, MS SQL Server and other nice stuff in the bundle (BIG savings!) It's easy to admin (not every ~30 employee company has a fully fledged IT department on hand) and it's relatively inexpensive too. There's no need for 2 servers here. Personally, I'd likely go for a ML110 G6 if I was buying HP. With the "basic" Xeon X3430 (quad core) and 4x2GB of DDR2. That should be enough to run everything you were talking about for the next few years no problem. Of course, you'll want to pick enough storage for your needs along with that (at $419 for as basic 750GB SATA drive, I'd be tempted to install those myself, especially if you want a few of them -- but then no warranty for them, and it may be hard to find a replacement down the road should one fail). Then again, not knowing what kind of data you'll be storing, we can't quite tell just how much space you'll need. RAID0 is not really an option here (IMO), RAID1 might not provide you with enough space (again, no idea what you really need, much less what you'll need in the coming years), leaving you only with RAID5 as an option (well, or no RAID at all, which might be just fine). Mind you, that might already be getting more expensive than what you were thinking. With the Xeon X3430, 4x2GB, the SBS 2008 Premium license (which includes Windows itself, Exchange, MS SQL Svr and more), four 750GB SATA drives in RAID5 (one online spare, one for parity) giving you 1.5TB usable, an internal tape backup (DAT), and 5 year warranty (4h, 24x7) it's already like $5000 (not including extra CALs). You could save a bit by foregoing the RAID5 though (it mainly means a longer restore from a drive crash, using tapes instead of array rebuild). Either ways it's not going to be really cheap. With 2 servers running Win 2008 and Exchange Standard and MS SQL Server Standard you'd be looking at about that much in software licenses and CALs alone (more if you want Oracle instead), hardware extra.

It totally is electronics at every level. The land of [V]HDL, FPGAs and all that "fun" stuff. And even if you go that way, the odds of working for Intel/AMD/nvidia on mainstream hardware are just about nil. There's very few jobs like this, and you have to be more than "just good" at it (and they're often in foreign countries). I've seen plenty of people who studied that stuff, but none who ever made it to such a position (last one went to work for ON Semi on stuff that would most likely bore you to death).

Not knowing how many users we're talking about (10? 100? 1000? more?) and what needs they have, we simply can't. DHCP and DNS are a complete non-issue. But Exchange-wise, you could simply need 10 small mailboxes (a few emails a week), or a few thousand boxes of several GBs each... That makes a LOT of difference for how much disk space and RAM you need. Same story for the "application server". Now knowing which actual applications it'll actually run (could be a simple php-based blog on apache, or a heavy app running on top of Oracle and what not at the same time), how many concurrent users (5? 5000?), and what kind of resulting load you'd have on the server, we have no way to tell anything. It should be somewhere between an old P2-class Xeon running SBS and a HP superdome-class super high end hardware with a high end SAN (multi-million dollar setup) There's just no way we could know with so little information. If you post way more details then we might be able to help.

There is no such thing as a "RSA" or "VA" accounts. The Administrator account is as much of an administrator as it gets (you're locked out of nothing). The installation (T-12 phase) is running as the system account though. Slipstreaming essentially means "integrating" (most of the time, referring to service packs). Just so you don't have to download and install a ridiculous amount of patches every time you install the OS on a machine. That's all there is to it. There is no such thing as a true dos-level admin account. There aren't any dos "parts" to begin with... It's likely getting redone, but for modern versions of Windows (Vista/Win7) and their brand new installer, not for OS'es that are a decade old and their even older installer (dating from the NT 3.x days). I don't see groups mentioned anywhere in his post, nor talk about installing software. He's talking about HKCU reg tweaks specifically, which at T-12 end up being applied to the default user profile, as the install is running under the system account. All new user profiles created afterwards (including the Administrator account) will inherit that (making it redundant to re-apply them to your account later indeed). I have no idea where you're going with the group policy stuff there either. It totally depends on the program and its installer. Some installers will provide options or switches to do this, others don't. Sometimes, it's merely the start menu shortcuts that are created in a specific users' start menu instead of the all users start menu (then it's only a matter of moving them) and sometimes there's a lot more to it (registry entries, files in the user profile, etc) in which case there is no easy way. It doesn't look like you're flaming at all, but rather that you got several things confused (accounts don't lock you out of anything, and don't magically give the Pentagon access to your computer -- sorry, we are not in a movie). But no, there is no universal way to install anything for all users, if that answers the real question. And patches are applied system-wide. And if you want to mess with the default users' registry (and not at T-12), you'll have to mount it first. Not that I'd be moving to an OS that's a decade old, several versions out of date, with a 0.5% market share, whose extended support is just about over, has poor compatibility with modern apps and all that...

You can't slipstream VS SPs. This has been discussed a few times before, like in this topic last month. and frankly, I have no idea why you'd want to use the old 2005 edition.

Here's another option, something I threw together in a few mins, quick and dirty: option explicit on error resume next dim shl, fso, pc, list, qry, wmi, colping, ping, rwmi, colcomp, comp, user, logfl const in_file = "pc_list.txt" const log_file = "caught.txt" Set shl = createobject("Wscript.Shell") set fso = createobject("Scripting.FileSystemObject") set list = fso.opentextfile (in_file, 1) '1=ForReading do until list.atendofstream pc = list.readline if(pc<>"") then process(pc) Loop function process(compname) qry = "Select * From Win32_PingStatus Where Address = '" & compname & "'" set wmi = getobject("winmgmts:\\.\root\cimv2") set colping = wmi.execquery(qry) for each ping in colping if ping.statuscode=0 then 'PC is reachable, verify IP range if(left(ping.protocoladdress,8)) = "10.10.6." then 'we've got ourselves a rule breaker! set rwmi = getobject("winmgmts:{impersonationLevel=impersonate}!\\" & compname & "\root\cimv2") qry = "Select * From Win32_ComputerSystem" set colcomp = rwmi.execquery(qry) for each comp in colcomp user = comp.username next set logfl = fso.opentextfile(log_file, 8, true) '8=ForAppending logfl.writeline(now() & ", " & compname & ", " & ping.protocoladdress & ", " & user) logfl.close shl.run ("shutdown /t 30 /c " & chr(34) & "dont do that" & chr(34) & " /m \\" & compname) end if end if next end function It will read the PC names from the file called "pc_list.txt" (no need to worry about trailing blank lines either). Then it'll "ping" them using WMI. If they're reachable and it starts with 10.10.6. then it logs everything (timestamp, computer name, IP and logged on user) in caught.txt (just change the file names in the constants) in CSV format so you can see who are doing it and finally calls shutdown the way you wanted it. It executes pretty much instantly (<1sec for 15 PCs here, ICMP ping latency being the biggest slowdown). No error-prone text parsing of slow-running utils (like ping or nslookup) involved either (which usually fails should any error message should be returned instead of the expected output -- expect those to crash for a number of reasons, like the PC being turned off, wifi glitches, DNS not resolving, an IPv6 address being returned instead, your own connection having a hiccup or many other common issues -- quite error prone really, in fact, the "solution" in post #7 doesn't work *at all* on Win7, it just hangs there, even with a valid host!). The only "external requirement" is shutdown.exe which you wanted to call. It would be trivial to log different stuff, matching bad IP ranges using regular expressions or whatever else you so please. It's not tested very much (only inside one VM, as my entire "real" network is all on IPv6), poorly commented, ugly in general, and has little to nothing in terms of error handling or anything like that. It assumes the account running the script (you, or whichever user account you'll use to schedule this to run every few mins) has permissions to run WMI queries on the remote PCs, NTFS permissions to write the log file and such, so you might have to do some debugging (run whateverscriptname.vbs //x to start the debugger -- visual studio works fine for this too) Hopefully that helps

Well, if you take it like that, then perhaps your $100 9600GT card is faster than the $60 card I picked. Or the $60 card I picked is much faster than some $80 9500GT's. See? It works both ways... Yes, we wouldn't want people to have a sense of humor around here. Jokes are bad. It's not like most people register to complain about drivers (for any hardware). I've seen plenty of people asking about BSODs, which we found out later were caused by nvidia drivers. These issues are fairly well documented (again, there were even blog posts from people like Mark Russinovich depicting them) I gave my last one away 2 months ago. So nope, not anymore! All my endless video issues are now solved No more of this error: no more of this one either: and H.264 decoding now works in hardware too as a bonus! Couldn't be happier. Thanks to nvidia for selling me an ATI card. First, you don't, unless you're running an OS that's from 2001 or before. And second, you still don't if you only want the drivers. Not that I even see this as a "bad" thing in the first place. If that's your biggest complaint, then it's not saying much. You're assuming most people whose computer fail because of this even have a clue about the cause. 99% of people who got G84/G86 issues will return their computer to be fixed under warranty. Either ways, it looks like you're trying to deny issues to which several large OEMs like Dell, HP and Apple openly admitted to. The point is, $60 is a decent price point for his needs, whereas $80 is pushing it for a lot of non-gamers (seriously, he's using a 8400 -- I doubt he's into all the latest games). Below $60 you get very low end cards that are no better than his existing card, around $60 you can get decent cards, and above that it's mostly gaming-oriented. The 9500GT is most likely fast enough for him, but the similarly priced 4670 is faster (*and* has better drivers). It looks like you're just trying to find a lot of excuses because nvidia doesn't have any good card at that price point (well, that would be most price points right now -- and no, I still couldn't care less if nvidia hasn't released new stuff yet). Either ways, I'm getting tired of even answering your always overly sarcastic first replies to any of my posts, like this time again (which always tends to degenerate like this): To a post where I was talking about 2 cards at the exact same price (suggested by someone else no less). Sp please lose the sarcastic tone next time. Either ways, the OP had several suggestions about video cards, so it's up to him to pick one now. This thread is clearly going nowhere now (and not helping him any further) so I'm closing it. If he has any more unanswered questions, then he can open a new one.

LOL. When do I start? I merely used that setup he stated out of lazyness (I was genuinely too lazy to price the 3rd option). Then again, all 3 mobos you listed only have 2 DIMM slots... But yeah, they're basically all decent options, and it'll be a gigantic step up from a P3 regardless of what he picks.

This is obviously homework, see the forum rules, particularly rule 2.c Consider this your first and final warning about this. [closed]

Well, there's not a whole lot of difference (building a system-wise). It works just the same. Let's compare both options (your initial and a decent AMD setup): GIGABYTE GA-G31M-ES2L LGA 775 Intel G31 Micro ATX Intel Motherboard - Retail Intel Pentium E5200 Wolfdale 2.5GHz LGA 775 65W Dual-Core Processor Model BX80571E5200 - Retail Kingston HyperX 2GB (2 x 1GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400) Dual Channel Kit Desktop Memory Model KHX6400D2K2/2G - Retail Total: $178 AMD Athlon II X2 240 Regor 2.8GHz Socket AM3 65W Dual-Core Processor Model ADX240OCGQBOX - Retail ASUS M4A785T-M/CSM AM3 AMD 785G HDMI Micro ATX AMD Motherboard - Retail Crucial 2GB (2 x 1GB) 240-Pin DDR3 SDRAM DDR3 1333 (PC3 10600) Dual Channel Kit Desktop Memory Model CT2KIT12864BA1339 - Retail Total: $204 So for ~$25 extra, you get: -a motherboard that uses a socket that hasn't been replaced already (newer, faster CPUs for cheap down the road) -far better onboard video: DirectX 10.1, it even does decoding of H.264 HD movies in 1080p in hardware (be it Blu-Ray, x264, etc), it also works great for aero glass, Photoshop CS4 OpenGL acceleration and the like -you get DVI and HDMI outs out of the box. Going Intel with a G41 mobo negates the slight price advantage the E5200 has (while still being slower), and so does buying a $50 video card... -you get a PCI Express 2.0 x16 slot -4 DIMM slots instead of 2 (more expansion room) -it uses newer and faster DDR3 (also means cheaper to upgrade down the road -- you'll easily get your $20 back then, and it'll be easier to find at that point too) -you get more SATA connectors, a toslink out (more analog too) and more USB at the back (and more headers too) -a newer chipset (785G is brand spanking new i.e. from 2009, whereas G31 is from 2007 and its ICH7 is from 2005) -you can get models with eSATA and firewire too for very little more etc. There won't be a problem dropping a very fast quad core on this setup in a few years, along with some more DDR3 and a fast GPU should you feel like it, and without breaking the bank (AMD already has quad core CPUs for $100) Just saying that it's worth considering, especially if you got the extra twenty bucks or so to spare. But it's not so easy undercutting the E3200 setup. Either ways, it'll run circles around the P3 (anything would)

Then again, Intel setups with DDR3 are still relatively expensive. And 99% or so of mATX boards for C2D's only have 2 DIMM slots, so if you wanted to move beyond 2GB, you'd have to throw away your existing RAM. For a few bucks difference with your current selection of parts, you could build a box with an Athlon II X2 240 CPU (has VT as well, same performance too), DDR3 RAM, and a LOT nicer motherboard. It would be a LOT nicer in many ways: onboard video that's a million times better in terms of performance, more video outs than just plain old VGA (often you get DVI and HDMI, sometimes DisplayPort too), 4 DIMM slots is standard on 99% of them, it would be using DDR3 that's now cheaper, and perhaps have a few more features like eSATA. I understand what rock bottom price is a big factor here, but LGA 775 is pretty much dead, DDR2 prices are on the rise (and you have no expansion capability either), and low performance GMA video with VGA-only output might not be what you're wishing for either. Just another option worth considering

That is the most apparent one perhaps. The 100MHz clock difference shouldn't really be noticeable (4% difference). FSB wise, it makes no real difference unless its BW starved (and it's not). That leaves the cache difference, which will have a varying effect depending on what one does, but then again, if you look at passmark scores, the E3200 gets 1458 and the E5200 gets 1329 (yes, lower!) so it can't be much slower (again, depending on your workload). The main difference for me, and what makes that I'd sooner buy a E3200 over a E5200 even if they were the same price, is that the E3200 has VT. That means you can run x64 guest OS'es under vmware/vpc/virtualbox and others, as well as use Win7's XP compatibility mode, none of which won't work with the pricier E5200.

Just like you can get them for $60. The point was, they have better cards at that price ($60). But hey, I'm sure you can overpay for any card (that seems to be your "point") You said popularity = being good, whereas there is little to no correlation. Seems like you didn't get it. Not that nvidia is impressively popular in the first place mind you. It's very much becoming one, thanks to Vista, Win7, OpenGL accelerated apps like Photoshop CS4 and others (2001 is over). It does matter to people buying new PCs and OEMs. Either way, nvidia still losing market share to ATI (and Intel) lately. Funny, because I could have swore for a second you said nvidia Worst drivers ever (their XP drivers are OK, but their WDDM ones are just starting to be usable now, like 2 years late; oh, and broken H.264 HD decode acceleration too... or does that mostly work now, 2 years late?) Sorry, they won't synchronize product launches for you, nothing I can do about that. You won't see *me* making excuses the day nvidia has best performance on their new cards that cost less (anyone could say "Hey, those are new. ATI will make new cards next!" too) Then look at this, for example the numbers in gray. Yep, up to double the framerate (e.g. in Stormrise). Not bad for a cheaper card. And yes, I'm sure vaporware is probably better (not that it helps him getting his money's worth out of a $60 card now) No, all that for saying a $80 card was better than a $60 card (or was it saying more expensive cards can be faster -- when ATI has faster cards at the same price?), then going on about that nvidia cards are less failure prone when it's anything but the case, and then somehow trying to say market share = being good (even when it even plays against yourself). Mind you, I had already mentioned a $60 nvidia card in my first post (something you haven't done yet, despite having written two lengthy posts )

Yes, funny how when you raise the price point, you can get faster cards eh? I was going for Zxian's suggested (and reasonable) price point, at which ATI has better offerings. A $80 nvidia card beats a $60 ati card, just like it's no challenge finding $80 ATI cards that are faster than a $60 nvidia card (actually, it's easy, even at $60). No that he likely needs that kind of performance in the first place. You mean their competition with record breaking G84/G86 failure rates? Right. So Intel (GMA) must be absolutely fantastic cards as it has the vast majority of it and are still growing. Much like McD's must serve great food because loads of people eat there. You just can't argue with that indeed. BTW, nvidia is losing to Intel and AMD lately, so that must mean bad things as well! I watched a wicked demo of a 5870 yesterday (triple monitors @ 1920x1200). Absolutely amazing. Oh, and its little brother the 5850 pwns a GTX285 too, despite being significantly cheaper (and also using less power)

That's not extreme. Some people manage to push it beyond 4GHz. That's not much of an OC (not bad for a P4 perhaps). My current CPU OC'ed from 1.8 to 3.4

With a good cooler (and a OC-friendly motherboard of course), you can get more like 3.5GHz. That would pwn the i7 920 @ stock speed (you'd be around the i7 975).