Log in
Sign up
View all  |  Close all
#
Subject
Posted by
Actions
1
Why are updates lost?
Rolf Bäckström
- Close
I complained several times before about the severly bugridden FTB Smart Matches process. Those bugs are numerous, so numerous that the code should be rewritten and I have not bothered to list them. There is one particular aspect, though, that deserves particular attention. In the Smart Matches you may update the name, birth date, death date etc. When the update concerns rarely encountered items like burial, wedding, occupation etc. you may take the trouble to update them, but do not bother, they do not update, they just seem to update!!

Yesterday I again took the trouble of going through some geographical discrepancies found in a Consistency check and carefully correcting them. The process contains some bugs and could be more transparent, but after saving the data and repeating the scan, the errors are still there as they were. Even after closing down FTB and the PC and reloading, errors still remain intact.

Both error types seem to be similar enough to be part of a greater error scheme, i.e. a common erratic algoritm. This error existed already a couple of years ago when I was using MS Vista and now when I am using MS W7. 

Have you noted this and do you intend to review the Smart Matches process and the algoritm problem any time soon?

Rolf Bäckström
Finland
- Reply
2
RE: Why are updates lost?
Marianne
- Close
We are aware of the fact that the system is not flawless.
As a result of these corrupted pools ( as we call them) a new mechanism for calculating Smart Matches is already in our developer's hands and on the road map as part of the Smart Match improvements. 
This new mechanism will reduce the number of false positive matches dramatically. 
It's much more robust than our current system.  I'm not making any promises about time frames for release of this new system - there is still a lot to understand about the scope of the project before we can define and promise a time - 
but I thought you would be interested to know it's in the works. 
Marianne
Netherlands
- Reply
3
RE: Why are updates lost? [2]
Rolf Bäckström
- Close

Many thanks for the news, although the epitet "flawless" is rather mildy put! Unfortunately my question was not formulated very well. Two independent issues were put forwarded and it was not completely sure which one was addressed. I conclude that you addressed the Smart Matches issue and that the other one was not addressed at all. It is the first time I have got feedback on my numerous complaints on Smart Matches and that somebody else but me and many fellow genealogists are deeply concerned. The Smart Matches feature signifies a major change in paradigm for the common genealogist. The idea is brilliant but probably your master plan was that Smart Matches would just nicely support conventional genealogy work, where you dig into databases and documents to find solid data. You could ascertain yourself that the data put into the family tree matched the original data. What now happens is that people indiscriminately and predominantly transfer data from other family trees using Smart Matches, without bothering to check whether the data is corroborated by solid evidence or not. Actually it is alleged that when somebody has put in something in his family tree, it is assumed correct. I have seen that when I have built up a branch in my family tree, where I am the sole provider and owner of the information content, I have received information from MyHeritage that now my information has been corroborated by four fellow web site keepers and may now be assumed correct. Being the sole originating information source I know this is an absolutely wrong statement! It is something called circular evidence, which is not evidence at all, but does everybody understand this? I see this as an example of many that the quality of MyHeritage information content is corroding seriously and is not under anybody's control.

A major effort should be put into error checking (methods, tools etc.) when using Smart Matches and the whole process should be carefully reviewed and redesigned. Also the impact of the Smart Matches on a much wider scale should be reviewed and addressed. 

Rolf Bäckström
Finland
- Reply
 
Loading...
Loading...