Wikipedia:Reference desk/Archives/Computing/2014 January 16

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.

January 16

Windows Server 2008: Indexing service running / not running

Hello there,

I've got a problem. I don't understand what is going on. I wanted to set up Indexing on WinSer 2008. There are good websites that walk you through the process. So, if I open my "Server manager" I can see in the leftmost pane "Roles," I click on that. I see "File Services" and in the central pane I see "Indexing Service cisvc Running." It really says RUNNING! In the pane for "Role Services 3 installed" it says. File Server - installed; windows server 2003 File services - installed; "Indexing Services -- Installed." Also if I click on Configuration==>Services and check Indexing Service in the center pane (large) I can see that "Indexing Service Started Automatic." It means it is RUNNING.

This is the problem. If I find "Indexing Options" Wizard via a START==>Search window I can see that the whole wizard is dimmed and on top of it it says: "INDEXING IS NOT RUNNING." I made it all capital, in fact it is in lower case.

If I try to search for any file I need, I can see that there is no indexing. Besides I've never set up folders for indexing which I should have been able to set up via "Indexing Options."

The truth is I downloaded Ubuntu, paid them fifty bucks and now I cannot find it in the directory tree.

What is the problem? Thanks. --AboutFace 22 (talk) 02:49, 16 January 2014 (UTC)

I just found this http://www.windowsnetworking.com/kbase/WindowsTips/WindowsServer2008/AdminTips/Admin/TroubleshootingIndexingisnotrunningonWindowsServer2008.html website It says I installed the wrong windows service and in fact I needed to install Windows Search Service. Is it true? --AboutFace 22 (talk) 03:13, 16 January 2014 (UTC)

Maybe I am missing something here, but I don't understand why you think you need to enable indexing if all you want to do is find your Ubuntu folder. Indexing speeds up repeated searches of files in commonly used locations, but I don't think it will speed up a one-off search for a single folder, especially when that search potentially has to span you whole file space. Gandalf61 (talk) 15:08, 17 January 2014 (UTC)

Indexing is very useful. You never know when you need it next time. I uninstalled Indexing Service as that website suggested, installed Windows Search Service and the indexing is still not working. It has become a big hassle. Now I went to my other OS: windows 7 (on the other hard disk), included the Windows Server 2008 hard disk in its Indexing and wait when it collects all indices. This might be the only way for me to search in the server. I wish someone with experience could comment on why I cannot set up indexing in the server. --AboutFace 22 (talk) 15:42, 17 January 2014 (UTC)

image threshold tool?

I want a tool to read two similar images and output a comparison image: white where image A is brighter than image B, black where image B is brighter than image A; don't care where they're equal. If ImageMagick can do that, what's the syntax? —Tamfang (talk) 03:46, 16 January 2014 (UTC)

I should mention that the images are monochrome in content, though RGB in format. —Tamfang (talk) 10:46, 16 January 2014 (UTC)

I'm wondering if GIMP might not be better for this. Image A can be positive and change image B to negative. You don't then need to know the syntax. One just merges the two layers together. For final adjustment, slide the mid scale pointer up and down (on both layers).--Aspro (talk) 14:11, 16 January 2014 (UTC)

To merge layers, the top one needs to have alpha 1/2, right? And I can't find a command to change (or examine) alpha for the whole layer. (I have GIMP 2.6.12) —Tamfang (talk) 22:55, 16 January 2014 (UTC)

Changing the alpha of an entire layer is easy. Just open the "layers" dialog, select the layer you want and use the "Opacity" slider. SteveBaker (talk) 03:39, 17 January 2014 (UTC)

Ah, it's hiding under the Windows menu rather than the Layer menu. —Tamfang (talk) 08:32, 17 January 2014 (UTC)

I think compare is what you want in ImageMagick.--Salix alba (talk): 15:44, 16 January 2014 (UTC)

Never mind, I can't get ImageMagick to work; seems to be missing a library. —Tamfang (talk) 22:55, 16 January 2014 (UTC)

Thanks all, it's partly moot. A belated bit of algebra told me I can get the effect I want with a cunning POV-Ray pigment and simple thresholding. —Tamfang (talk) 08:32, 17 January 2014 (UTC)

What does the limit of 32 or 64 bits actually imply?

The maximum representable value might be 4,294,967,295 or 18,446,744,073,709,551,615 for 32 or 64 bit computers. But can't you represent bigger integers than 18,446,744,073,709,551,615 on a 64 computer if you simply break up the integer into two and save it in two registers? I see that this will be slower than one integer in one register, but it doesn't have to mean that strange mistakes will happen. In the same way that I know that if an odometer was at 999,999 and I drove 2 extra miles, now the car has 1,000,002 miles not 1. OsmanRF34 (talk) 18:03, 16 January 2014 (UTC)

Certainly integers to large to represent with 32 or 64 bits can be dealt with by breaking up into pieces. This is done in some cases. But consider a statement like this in a typical programming language:

A = B + 227;

Ordinarily, the complier will be designed so that A and B must be limited to values that can be stored in the computer's registers, whatever size that might be. So if larger values are to be manipupulated, the programmer would have to test each number to see if the result will be too large, and write code that explicitly carries the excessive result into a new variable. Alternatively, the programmer could find and use a subroutine library that handles larger-than-usual integers. Jc3s5h (talk) 18:23, 16 January 2014 (UTC)

I don't think "ordinarily" is right here. In Java, the sizes and formats of the integer types don't depend on the CPU type. In Python, integer operations that overflow the machine register size automatically upgrade to arbitrary-precision integers. In C and C++ the size of int may differ, but not necessarily – it is normally the same size in 32-bit and 64-bit Windows. Perl and Javascript's "integers" are IEEE double precision floating point, which can exactly represent integers up to ±2⁵³ on 32-bit and 64-bit platforms.

Regardless of the language, if A and B are 64-bit integer types, the compiler will do 64-bit math using whatever CPU facilities it has available. It won't truncate 64-bit results to 32 bits just because that's the CPU register size. The only thing you have to worry about is whether generic type names like "int" map to 64-bit integer types or not. Almost all languages that have 32-bit integer types also have 64-bit integer types on 32-bit platforms by some name or another.

Historically the number of "bits" of a system has been as much about marketing as technology. The Super Nintendo and Sega Genesis / Mega Drive were both called 16-bit even though the Mega Drive's CPU had 32-bit registers and both machines' CPUs had a 24-bit address space. For that matter, the current AMD64 / Intel 64 / x64 architecture has a 48-bit address space, though pointers are always stored as 64 bits. It's probably better to talk about x86 and x64 than "32-bit" and "64-bit", since they're better defined terms. -- BenRG (talk) 21:46, 17 January 2014 (UTC)

The major limitation of 32-bit systems is the storage of a memory address in a register. This makes it difficult in general purpose programming to use more than 4 GB of RAM - and in practice less is available to a program because the operating system and other programs will use some of the available space. Memory segmentation is used on some systems to allow programs to address more memory than the register size allows, but this can complicate programs and prevent use of very large structures (probably, this was more of a problem with the 16-bit processors; see x86 memory segmentation).-gadfium 20:16, 16 January 2014 (UTC)

Look at BigNum, the article there discusses using numbers larger than a 32 or 64 limit. RudolfRed (talk) 02:59, 17 January 2014 (UTC)

There isn't much problem programming a computer to calculate to more precision than it's underlying word length. HERE, for example is a web site that allows you to do calculations with up to 226 million digits (which is probably due to a memory size restriction rather than an actual computational limit). It's just like when people who've memorised the multiplication tables up to 10x10 can multiply numbers like 123x456 by calculating 456x100 + 456x20 + 456x3. "long multiplication". The difference is that it's much slower to do that than if you'd memorized the multiplication tables up to 1000x1000 instead.

SteveBaker (talk) 03:35, 17 January 2014 (UTC)

If your odometer was at 999,999 and you drove two more miles, it would read 1,000,001, not 1,000,002. Perhaps working with large numbers is not as easy as you thought? Matt Deres (talk) 13:44, 18 January 2014 (UTC)

Perhaps it's a 9s' complement odometer [i.e. like 1s' complement but for base 10]. :–) --50.100.193.107 (talk) 21:19, 21 January 2014 (UTC)

program that runs on xp only

not in the following os, not even in compatibility mode.. but xp is discontinued.. what can i do.. thank you --82.58.163.72 (talk) 21:10, 16 January 2014 (UTC)

You can use Windows XP Mode if you have Windows 7 Professional, Enterprise or Ultimate. You can also use VMPlayer. [1] --  Gadget850 ^talk 21:47, 16 January 2014 (UTC)