Just a random post: There were a couple of topics that I was surprised don't seem to have come up as projects.
1) Improved tools for new family search (maybe it isn't set up to be possible). I know that in cases where the work was already done but the card that was submitted didn't contain much information about the individual it is difficult to find that work on the existing new family search (don't get me wrong, I think the new family search is fantastic). Also there isn't a way to search by date range, one has to go to the old family search to do that.
2) OK, this is an oddball, but I have a long standing interest in development of software that recognizes and culls sleaze. I know this is an insanely difficult problem (which is why current internet filters perform so ridiculously poorly). Yes there is the K9 (or whatever its called) internet filter which can be used for home use. I was looking for something that uses image parsing rather than URL checking (which is I believe how effectively all filters work, just because image parsing is so unreliable in its present state)
"Missing" projects
-
johnmortal
- New Member
- Posts: 2
- Joined: Fri Oct 23, 2009 10:24 am
- Location: USA
-
JamesAnderson
- Senior Member
- Posts: 773
- Joined: Tue Jan 23, 2007 2:03 pm
On number 1, that would be a good feature request, just send an email to support and indicate it is feedback/suggestions, and I'm sure they will look at the idea. After all, with some names, there's no record in the older FamilySearch site for some names now.
On number 2, K9's parent company does use a parser that may be taking into account images, and it is said to be very effective at capturing and rating content that is image-based, multilingual, etc., and that even before their hand raters find out about the site from a user submission or enough hits. So far, it is very good at catching obscene content before you hit into it by mistake.
It is not completely foolproof yet, but is very close, and although it is very effective, they are working to make it better. So it is going to be possible to have an image parser but it may be a little bit shaky for a while until the programming and technology get better.
On number 2, K9's parent company does use a parser that may be taking into account images, and it is said to be very effective at capturing and rating content that is image-based, multilingual, etc., and that even before their hand raters find out about the site from a user submission or enough hits. So far, it is very good at catching obscene content before you hit into it by mistake.
It is not completely foolproof yet, but is very close, and although it is very effective, they are working to make it better. So it is going to be possible to have an image parser but it may be a little bit shaky for a while until the programming and technology get better.