When I try to import an opml file into scrivener, all the accentuated chars (é,à, ù, etc…) are replaced by esoteric code such as :
original text :
"En France, on estime que 5 millions de personnes sont concernés par le problème de l’alcool dont 2 millions dépendantes. "
imported in Scrivener :
“En France, on estime que 5 millions de personnes sont concernés par le problème de l’alcool dont 2 millions dépendantes.”
I try to import by drag and drop or by the file menu, same result.
I explain how it happens : the opml file comes from a mind map created in NovaMind 5. I use the command “export to opml format”
First I thought the problem came from NovaMind but I tried to open this file with OmniOutliner and the opml file opens properly with no chars corrupted.
So can you help me to resolve this issue, if Scrivener is concerned of course?
My config :
Mac OS 10.7.2
Scrivener 2.2
NovaMind 5 express 5.1.0
Scrivener’s OPML import is very basic - it expects UTF8 text which would contain unencoded accented characters. However, it sounds as though NovaMind is escaping all accented characters - Scrivener currently won’t translate these, I’m afraid; it’s just a limitation in the OPML import.
I don’t know if it would work, but you could try opening the exported OPML in TextWrangler and saving it in a different encoding, or doing a find and exchange for all the “incorrect” codes. Bit of a faff, but might be the best you can manage. I’ve often used TextWrangler to accomplish things that I wasn’t supposed to be able to do … It’s free, as well as being useful.
I’ve made a note to look at decoding these characters in a future update, too, although I need to look at the XML specs to check that they should indeed be encoded.
All the best,
Keith
Strictly speaking of standards, that should depend upon the declared character set at the top of the XML file. If the file is declaring as UTF-8, then there is no reason to encode high bit characters, and Scrivener is doing the right thing. If it is declaring a smaller encoding like Latin-1 or ASCII, then characters should all be encoded in the file and thus decoded on read.
At the moment Scrivener is ignoring that declaration (assuming UTF8), however, it does turn out that Novamind is declaring UTF8, so the characters shouldn’t really be escaped.
As Keith pointed out this seems to be a limitation in the Scrivener import. NovaMind escapes accented characters and Scrivener doesn’t seem to be importing those correctly.
We have tested importing a NovaMind generated opml file to OmniOutliner which is a widely used outlining tool on the Mac and it imports the file correctly. Our engineers say that escaping characters is in our scenario the better approach as we would otherwise need to special case characters such as <, > or "
Unfortunately it would not be a simple change for us to modify the export to deal with this limitation and because other applications seem to support the currently created OPML file, it seems that this might be better addressed in the Scrivener import instead.
Maybe you can contact Scrivener about this. If they have any questions or comments about this they can contact us directly at support@novamind.com
Regards,
Sam
NovaMind Support Team"[/i]
I don’t know the english equivalent but in French we say something like " to return the hot potato". *
In the mean time, I 've found a temporary solution : I open the exported OPML novamind file in OmniOutliner, I save it in .oo3 Omni proprietary file format and then in OmniOutliner I export it in OMPL and bingo when I import this file in Scrivener, the accented characters are all there.
*