Re: [Gramps-devel] data import + handle = augh

[Jérôme <rom...@ya...>]

Michiel,

> So what I am experimenting with is just a gramps-xml file with a <?xsl-stylesheet type="text/xml"?> processing instruction at the top. 

There is some security issues by using code on client side (like 
javascript or xsl). I like using this method on local but will never 
host this on the web !!!
Should be done by servers, no ?
	
If this is the philosophy of Web 2.0/3.0, I become too old (30 years!) 
:-D :( :'(

> the person/application

we need more memory, no ?
Also, Not certain that M$ will use a "standard" parser for IE ...
Furthermore, this looks like activeX :( :'(

Why not a prism application ?
http://developer.mozilla.org/en/docs/Prism




Michiel Nauta a écrit :
> Hi Benny and others,
> 
> I like to come back to the issue I raised the previous weekend, now that 
> I have a test-case online (at 
> http://www.home.versatel.nl/m.d.nauta/stemkohier/Oostdongeradeel.xml).
> I understand that within Gramps the calculation of handles is such that 
> it is perfectly save to import data, also from others users of Gramps. 
> But I am thinking of a different use case: offer genealogical data as 
> gramps file, so that data entry becomes something close to 
> drag-drop/import. So what I am experimenting with is just a gramps-xml 
> file with a <?xsl-stylesheet type="text/xml"?> processing instruction at 
> the top. So the casual visitor sees some html-like presentation of the 
> data while the thorough visitor can see and copy the source. For an 
> example, I would like to point you to the above mentioned file which is 
> just a list of old farms with their location displayed on an old map 
> (use "view source" and/or shift-click on a circle/farm and choose "als 
> XML" to see the Gramps-xml (the latter doesn't work yet in Opera).)
> 
> Now your argument of low probability of coinciding handles isn't valid
> any more, in view of all the villeins on the net who's joy it is to ruin
> other peoples digital lives. They can just read the handles that are 
> used, set up a similar website, trick people to also import their foul 
> data and thus overwrite genuine content. (Yes, at this point I do wonder 
>   if I am going paranoid; if so just tell me.) If Gramps, on import, 
> stamps its own handles on the data (and what about the changetime?) then 
> this scenario couldn't happen.
> 
> Another benefit would be that now, the person/application offering 
> gramps-xml content needs to come with good handles. The 
> Oostdongeradeel.xml file mentioned above, doesn't, it is just a random 
> number under 1 million, which is enough for the limited set of 
> primary-objects that are in this file.
> 
> Michiel
> 
> 
> Benny Malengier schreef:
>> Yes Michiel,
>>
>> that is the case, it has worked like that since I joined the project, but I
>> heard somebody say it used to be different longer ago.
>>
>> Anyway, the changes of handles being the same is extremely low, as they are
>> made behind the scenes in GRAMPS. So GRAMPS can use it as the _UID field in
>> custom GEDCOM.
>> So, in the future we could provide automatic merge on import based on equal
>> handle, usefull for collaboration. You should hence not change handles made
>> by GRAMPS, instead you should use the gramps id for whatever you want to do.
>>
>> At the moment however, merge-overwrite on import is a side-effect of how it
>> is coded, and is not a supported thing, but I can see it supported in the
>> future. Unfortunately the code gives no warning that it happened. In 3.0 I
>> added a warning when merge happens, but it still 'just' happens and destroys
>> much things you need to correct manually...
>>
>> Benny
>>
>> 2008/2/29, Michiel Nauta <m.d...@he...>:
>>> Hi,
>>>
>>> I was thinking of putting some data on the net in Gramps-xml format. For
>>> this I need to create handle attributes to tie all primary objects (or
>>> whatever <person>, <source>, <event> etc are called) together. To my
>>> surprise I find that the handles I created for the import XML file, are
>>> the same as the handles that I find back when I export the imported
>>> data. To my horror I find that if I import something with a handle that
>>> is already in the database, the old data is overwritten. I have always
>>> assumed that on import, new handles would be attached to the primary
>>> objects. Are there no security implications for not doing so?
>>>
>>> Sorry, I am still using Gramps 2.2.8 (I know I need to upgrade).
>>>
>>> Michiel
>>>
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Gramps-devel mailing list
> Gra...@li...
> https://lists.sourceforge.net/lists/listinfo/gramps-devel
>