torsdag den 9. juli 2015

Geni vs. MyHeritage

When looking for a generalogy program, you may have encountered either Geni or MyHeritage. Opposite normal programs they both work in a browser, which makes it easy to share data and collaborate.

They are both owned by the same company and have some sort of integration between them, but are basically very different.

Geni

Geni is easy mass collaboration with no privacy. If you have entered an person Geni thinks someone else has entered it will tell you and you can merge the trees, so you get a lot more people. If you get merge into the big tree, Geni can show how you are connected to millions of people. Further more other people can change in your.

Unfortuntely the merge features requires a pro subscription which is expensive. Most of the matches Geni finds turns out to be non matches. I usually say that I have used about 1 of every 100. That also means that other people very easily can merge wrong trees, so you really cannot trust it, if you suddenly is connected to Leonardo Da Vinci.

Using Geni you can pay even more. It can find profiles and records in MyHeritage, but even with the expensive pro subscription you have to buy an equally expensive data subscription. The matches from MyHeritage are much better than Genis own, but you have to copy all data by hand to the Geni tree, which is extermely time consuming.

Geni does not allow you to import existing data, but you can export 4 times as many profiles as you have entered. Some years ago they posted that they were working on it, but it seems like any development on Geni has halted.

Because Flash is on the way out, and that is used to show and edit the tree on Geni, they have started a version coded in HTML 5. It have not improved for the two years I have tried it and it definitely does not have the same finish as the main view.

I have used Geni for about 2½ years and has had both the pro subscription and data subscription.

MyHertiage

MyHeritage takes it the other way: You create your own tree and others cannot change it. It also has a match feature but it allows you to copy profiles into your tree. If you want more than 250 profiles you have to pay and it is not that cheap.
You also have much more control over who can see your work, so only you or your family can watch.

MyHeritage allows you to both import and export your data using the common gedcom format. It also has a program for Windows computers that can synchronize with the online account.

I have used MyHeritage as a secondary site for about 6 months, and has used it with Genis data subscription for about 1½ year. I have not had the subscription for MyHeritage itself.

Conclusion

Both Geni and MyHeritage provides online collaboration. Geni is for the open and MyHeritage is for the private. They are both expensive and there are much cheaper solutions.
They both provide a way to leverage other peoples work, but beware of errors and note that data you get without a source must be checked unless you want a lot of errors in your tree.

lørdag den 3. maj 2014

Amazon S3 curl upload: Policy Condition failed: ["starts-with", "$Filename", ""]

For some time, I have been struggeling with fixing a bash script that uploades files to Amazon S3. It have been working, but I got an error when the configuration of the service was changed:

<?xml version="1.0" encoding="UTF-8"?>
<Error><Code>AccessDenied</Code><Message>Invalid according to Policy: Policy Condition failed: ["starts-with", "$Filename", ""]</Message><RequestId>ED7D91115863CB7B</RequestId><HostId>mOQWlKRPdgeVeRJFStqt5N0Jcc/0cD1wqDRSiQUgaw87K7Xg/D5hqWG+zIw8ldDS</HostId></Error>Done.

So what did it mean. Searching Amazons documentation on http://doc.s3.amazonaws.com/proposals/post.html for starts-with provided the answer: A requirement had been put in, that the Filename field I send should start with "".
Why start with an empty string? The doc says, that it is a trick to ensure a field is filled out. So I did not have a Filename field - I simply added it and voila!

Now how does my command look like: I have logged in (the site requires that - the method is different from site to site) already obtained the keys needed for Amazon, which are embedded in the site. The general magic is:

curlargs="-b cookies.txt -c cookies.txt --silent --output $res"
curl $curlargs \
-F "acl=private" \
-F "AWSAccessKeyId=$AWSAccessKeyId" \
-F "Policy=$policy" \
-F "Signature=$sig" \
-F "key=$key" \
-F "success_action_status=201" \
-F "Filename=$fn" \
-F "file=@$fn" \
http://some-service.s3.amazonaws.com/

All other fields, except the $fn which contains the full path to my file, are obtained from the site by screenscraping. The hard part is getting the proper fields and the order correct.
Hope this will help a soul or two: Please comment if it does and please add knowledge.