Using a Desktop Search through the network

[Writer]

In our community-run library, we have about 20 computers connected through a LAN. Between them, we have upwards of 150,000 files stored on four computers. These files are of format .doc, .html, .pdf, .txt, .xls and .msg. We've been trying to come up with a search solution that will allow users (who are decently computer savvy) to search all the files, and them access them over the network.

Is it possible to:

• Run a freeware desktop search on the central server and access the search engine on the local LAN?
• Index network drives of four computers, and make them searchable?

We tried to implement this using Google Desktop with a web-server called DNKA (http://dnka.com/). Unfortunately, the latest version of Google desktop that DNKA 0.49 supports is GDS 4.2006.0627; we need to use GDS 5, or later.

[mouser]

great question.. looking forward to hear what suggestion people here come up with.

[lanux128]

if your drives are in NTFS format you can use Everything to search a networked computer. the instructions are in the FAQ at the website. btw, i just happen to be reading the FAQ but haven't tried yet.

[Shades]

Since you are a library, maybe it would even be smarter to use your own search engine. IBM and Yahoo were so kind to give their Omnifind product away for free.

With your own search engine not only the (more) technically savvy users can access your data but everybody...which is kinda the main purpose of a (public) library.

[davekelly]

My favourite is Copernic, which easily handles local and networked content.

[MerleOne]

There is a small tool (commercial) from Valixsoft which is supposed to do quick net searches (maybe not within files).  It is regularly offered on GiveAwayOfTheDay.  It's called Netsearch I think.

I would also recommend ScanFS from Saleen Software.  GUI is a bit overwhelming but it has many features and perform non-indexed search (slow) within files.

[Darwin]

dtSearch is another option - but, like the ValixSoft offering, it's pricey. Works well, though (I've got a copy and have used it numerous times over my small wireless network.

Archivarius might be worth a look as well...

[Darwin]

PS Software995 also offers SearchWithin in a freeware (ad supported) version:

The free version of SearchWithin will display a sponsor page in your web browser each time you launch the software and every fourth search. If you would prefer not to see sponsor pages, you may download a license key for $9.95. SearchWithin never transmits any personal information or query information. Please see our FAQ for details about our software license or to contact us.

[wperkins48]

Haven't seen anyone mentions it, but Index Your Files is my choice - by far. [www.indexyourfiles.com] Download link: http://www.freewaref...a_program_32912.html

I am indexing Four Drives on the WAN, plus my local hard drive  --  a total of over *1 million* files. Program creates five separate databases that I can mix-and-match as I wish when searching (search Local C Drive and Network K Drive, for instance, or just one drive, for another instance).

I update the index each day at 8:30 AM  --  and really don't notice that much of a performance hit at all. The index updating over the WAN is really unnoticeable since it is not local and I really don't notice when it is updating the Local Drive.

Takes a *long* time to search for text in files, particularly if your database is large, like mine, because contents are not indexed (Duh...). Searching for just files is lightning fast:

AND, you can preview some files in a Preview Pane within the program. So far, this is the pass/fail list on "previewing" (results may also depend on the "viewers" I have enabled, so your mileage may vary...) in my testing:

PDF - Pass
XLS - Fail
DOC - Sort of Pass. Can view text.
DWG - Fail
JPG - Pass
BMP - Pass
GIF - Pass
PNG - Pass
MP3 - Pass
FLV - Pass
MPG - Pass
AVI - Pass
WMV - Pass
ZIP - Fail

As you can see, I like the program...

[LuckMan212]

to the OP:  I'm very curious what solution you ultimately settled on (if any).   I looked at the suggestions and at least in terms of the freeware offerings, it appears that 'Everything' with it's ETP server option was a good choice.  However, it does not index the contents of files, only their names.

[vixay]

I use Locate32 for the same. Though it only helps for one user... you could setup something complicated by deploying to other computers and accessing the database from a central location. Had something like that done for a client. You can check out their forums for some nice networking help as well. Maybe one day i'll write a mini-review about that, or a network searching article.
Anyway try it out, and maybe a less hassle solution would be better for you.

[aenache36]

Greetings.

If you need:
1. free software
2. able to index usual file types and mail(Outlook mainly)
3. able to index web sites
4. able to index network mapped drives
5. able to handle vast amount of documents
6. (very) short learning time
7. the possibility to use the computer all the time(X1/Copernic users know what I am talking about...)
8. client - server(excellent for LAN/ corporate use)
9. Boolean syntax supported, proximity/fuzzy search included
10. Open source/ OS independent(more or less)
11. superb visual display of the results

you might take a look at one marvelous product called AutoFocus, more info you can find at ADUNA.

HTH

[peter moore]

or you check this one: Lookeen
it´s a outlook-searchtool but it also can be used for normal file-types and it even searches within this files...
it works only with outlook (could be a little disadvantage, but at work, outlook is used the whole day, so this won´t be problem)

greets

edit by jgpaiva: removed link

[jgpaiva]

peter moore: I believe you're trying to push Looken. It is not adequate for the original question, as it can't search files a whole computer, let alone a few computers connected in LAN.

Of your four posts, three are about looken. Thus, I can only conclude you're trying to advertise this product even in situations where it doesn't make any sense. Furthermore, you're not stating your affiliation with looken, which goes against the forum's rules.
It's a dishonest practice, and I wouldn't buy software from a company which works like that. Probably if they're bringing in users by fooling them, they aren't producing quality software.

[Stoic Joker]

20 machines running (constantly reindexing) a "desktop" search across the wire?!? Zoikes!

When letting the Xerox rep talk us into using their ScanFlowStore software for our paperless service ticket system I was actually stuck with the exact same question. It seems Xerox (the SFS people actually) had spent their time creating a truly flexible metadata based document indexing system ... and just not gotten around to doing a search app just yet. The rep suggested we use Google Desktop, which I politely told him was a totally retarded idea.

What I did do was leverage the completely included with Windows (and fairly customizable) indexing service to create a single index catalog on the file server which can be searched in either a plain text or SQL-esq fashion using the Query the Catalog part of the Indexing Service MMC snap-in. I did end up having to write a small search GUI for the office staff to connect to the remote catalog but the system has been working perfectly for about 3 years now.

[vixay]

that's interesting. How fast is it compared to the desktop search programs we have? I have heard that the indexing system by ms slows downs things a lot!
Care to share the GUI for your search app? do you just specifiy an IP address and use the logged in users credentials? or is that not necessary for accessing the index?

how about integration with the intranet?
just idle questions.

[Stoic Joker]

Ironically I almost never use it, but the office staff does and they haven't had any complaints (which is odd for an office staff...). But for a test I queried a list of all of last years service tickets for client X, and it took about 2.5 seconds to pull the 20 test client tickets from the 7,000 other tickets.

Unfortunately I don't have time to strip the company identifing stuff out of the code. But I originally started it with a C# code sample/demo I found at www.CodeProject.com - Currently logged in user creds are used, & you can connect to and query either a local or remote catalog db.