tpersky@suntory.dcrt.nih.gov (Ted Persky) (08/14/90)
It has come to my attention that our staff on the whole spends a great deal of time during the day searching for that one certain file in a large directory tree. For example, try finding a certain fragment of source code in the X11 distribution. The people in our lab always seem to be asking each other to help them locate the path name for "foo.c". What I'm wondering is whether anyone knows of a tool where one can define a directory as being the root of a "large file tree" and have an index of some sort placed at that root. After that is created, each person who creates a file in that particular sub-tree would type in some sort of librarian command to create an entry in the index with a brief description of the file. Then the index (in database form, preferably) could be queried to locate the path name for a desired file. This would be ideal if people such as MIT could create this for their distributions of X, or UNIX vendors for their source distributions. If you could send me e-mail concerning this, I'd greatly appreciate it. Thanx, Ted Persky phone: (301) 496-2963 Building 12A, Room 2031 Internet: tpersky@alw.nih.gov National Institutes of Health Bethesda, MD 20892
hartman@ide.com (Robert Hartman) (08/15/90)
In article <335@nih-csl.nih.gov> tpersky@alw.nih.gov (Ted Persky) writes: >It has come to my attention that our staff on the whole spends >a great deal of time during the day searching for that one >certain file in a large directory tree. For example, try finding >a certain fragment of source code in the X11 distribution. >The people in our lab always seem to be asking each other >to help them locate the path name for "foo.c". If you have the BSD version of find, you can use the "fast-find" feature to build the database you want. This wasn't documented very well in the BSD man page. However, if you can get at a copy of a Sun 4.x man page, it's documented there. -r
stluka@software.org (Fred Stluka) (08/17/90)
In article <335@nih-csl.nih.gov> tpersky@alw.nih.gov (Ted Persky) writes: > The people in our lab always seem to be asking each other > to help them locate the path name for "foo.c". > ... > What I'm wondering is whether anyone knows of a tool where > one can define a directory as being the root of a "large file > tree" and have an index of some sort placed at that root. After > that is created, each person who creates a file in that particular > sub-tree would type in some sort of librarian command to create > an entry in the index with a brief description of the file. > Then the index (in database form, preferably) could be queried > to locate the path name for a desired file. This would be ideal > if people such as MIT could create this for their distributions > of X, or UNIX vendors for their source distributions. Can't help you with the solution you recommend (an indexer), but if you are open to other solutions... Teach the "people in the lab" about the "find" command. You can even make it more convenient for them, by defining an alias: alias dirr 'find . -name \!* -print | sort' which allows them to cd to the root of the tree where the file is known to reside and type: dirr foo.c Encourage them to use it on the smallest tree they know to contain the file because the search is slow. --Fred Fred Stluka Internet: stluka@software.org Software Productivity Consortium UUNET: ...!uunet!software!stluka 2214 Rock Hill Rd, Herndon VA 22070