"Jayjay" <jj****@notmail.com> wrote in message
news:40***************@news.cis.dfn.de...
On Mon, 12 Jan 2004 15:25:32 +0000 (UTC), "Fletcher Arnold"
<fl****@home.com> wrote:
The key to getting anything useful out of this system would be to store
someproperly formatted data at the time you first receive the document. Sure
you could write some code which scans for key words but that is not very
sophisticated and, since resumes for people in the IT sector often
containvast lists of software / technologies with which they are supposedly
familiar with, I'm not sure it's of much help. Perhaps it's because of
thissort of automated search by employment agencies is so common that
applicantsfeel the need to write such long lists.
If you gave examples of the sort of searches you intend to do, someone
mightbe able to offer better advice. If you simply wanted, for example, a
listof any people whose resume mentions the term 'visual basic', 'vb', or
'vba'then perhaps this is OK but my guess is that the list would be pretty
longand not very useful. If you wanted to find someone within 25 miles of
London with at least 5 years' database development experience, including
project management, with German language skills being an advantage, then
youmay well wish that you had read (or at least skim-read) these resumes and
entered some key details into a proper database structure.
Alternatively, use the established agency techniques of trying to match
candidate to employment without anyone actually reading any resumes and,
ideally, without reading the job requirements from the client.
</rant>
Fletcher
Point well taken. :-)
These resumes are for biological research or medical research type
companies.
I've asked the guy to give me a list of the technologies he
specializes in. The database of his company contacts is broken down
as such
3 company types. Biological, Medical and Diagnostic
Areacode for phone number. then State. for contact information.
Then he wanted a search for technologies available. Technologies can
be things like molecular research, cellular division, etc (he's
compling a list).
This database is a contact list of companies to market the candidates
to. So its backwareds. He gets a resume and then goes in and does a
search on technologies to get a list of companies that this candidate
might be good for.
But, he gets his resume's from candidates in Word format. (he
requests them to come in that way). He would like a way to store the
files in a directory, then run a search on the text inside the
document for some of the key words - to be able to backwards market
certain companies.
He didn't really want to put the word files into his database as
candidates come and go and he was trying to keep away from the
maintenance of cleaning up the records. Also, he prefers the paper
trail and having the full resume on hand in front of him so he can
read it. :-)
Then again, maybe a simply thing would be to have the database setup
that when he opens the word file, he runs a macro that selects all the
text, and copies it, then pastes to a memo field in the database.
That way I can have the database search the text in the memo field.
(The guy is breaking out on his own as a recruiter. so he doesn't
have the funds at this point in time to purchase the recruiting
software that is marketed out there. He's my husband's best friend
and a great guy, so I'm just trying to give him a hand here).
Those are certainly reasonable ideas - and I would agree that keeping the
Word documents is important. I was simply suggesting that you could add
some more structured data in addition to the actual Word document.
I am less convinced that cutting and pasting from the document is a good
idea, since individual resumes can contain elements which won't conform
nicely to a plain-text memo-field. At best the discrepancy might only be a
few funny-looking characters, but if they have created their name with a
Word Art object to make the document look nice, your plain-text memo field
will obviously miss this.
It depends how much time you have to build this, but you could write an
import routine that might attempt to do a better conversion than cut and
paste. For example, whenever a new document is received you could save them
in a standard place like "C:\Resumes\Inbox\" the import routine would then
create a record with some basic input such as name, surname, etc and other
fields being completed automatically (such as DateEntered). The document
would then be renamed to a standard name like a 6-figure format of the
RecordID and some form of 'automated cut and paste' would be done. The file
could be moved to "C:\Resumes\WordDocs\000156.doc" and this path would be
stored with the database record.
Whatever you decide though, having data in the tables (for speed and
structure) and retaining the original document is the way to go. I'm sure
there is no software which offers a magic fix and provided you are sensible,
I don't see why you couldn't hand him a perfectly workable solution. Good
luck.
Fletcher