I am not quite comfortable uploading all that info and media --plus it would be over 15gb.
I will try to find test tree that fails to do that when time permits. But I agree the server I am hitting (or other users) is a possible cause since It does not seem related to computer or account )
[EDIT] in all the test cases I the media option was NOT checked
I wonder if this is simply due to the communications protocol supported by the API. What level of error detection and correction does it have? What does it do when there are lost packets? This is not a banking, financial trading, national security application so the comms protocol may not be a particularly robust and reliable one, because that adds overhead, is slower and costs more.
Why did MacKiev find it necessary to introduce the Internet weather indicator for Family Tree Maker TreeSync? Red=Don’t, Orange=Risky, Green=OK. Same API. You heard lots of stories of sync problems on it and, presumably, that has been mitigated by giving users a traffic light.
I wonder if it is a streaming protocol that just ignores dropouts and carries on. Recently, I’ve been working with a TreeShare downloaded database from someone else that Renee had described as corrupted. That was suggested by an Integrity check which turned out to be due to the MacOS vs Windows differences in RMNOCASE but I also noticed that there were almost a thousand custom fact types with names of people or places. I doubt that they exist as custom fact types on the Ancestry Tree but might come about as a result of the loss of a block of the stream that mashed up some XML (?) tag’s contents with others.
That’s just a guess by a blind philosopher feeling his way around the elephant!
Okay I did more test.
used US Presidents GEDCOM and imported into 2 diff RM databases.
test #1 no changes was as is – uploaded to Ancestry – no issues all 100% people successful.
test #2 added a custom fact “ClassFact” added note ending “Class,” also sentence ended in “Class.” – all people successfully uploaded to Ancestry.
so explain that – based on what I understood you said @rzamor1 most or all should have failed.
@TomH any thoughts?
I’m not understanding why you continue on this tangent. This “angle” is not helpful to your friend’s problem. They said everything worked fine the 2nd time and that there was no “class.” in their data.
because clearly there is an unknown problem between RM & Ancestry – possibly with API but we only know it seems mostly random. I can not
replicate any scenario to fail consistently.
That’s clearly not what I meant. Badgering Renee about “class.” is not moving the ball down the field.
I was not attempting to badger Renee.
I was not the one whose raised issue that ‘class’ would cause failure (it consistently does not in my test). I do not know what Renee meant, knew etc. – I only know what her exact words said. She did not correct what she said. Moving on from ‘class’. …
It is possible the true root cause might be near impossible to determine, but until something is proven with a high degree of certainly that reason should not be blamed. One should avoid the user being the root cause of issue also. The goal would be to identify issue so they can be fixed. I understand resources are limited to determine root causes. (Worked at college as Blackboard and LMS admin – so there was just me).
