Could not save a blog entry on Blogspot

beogl

Dear Zotero,

I just tried to save a Blogspot blog entry to you, see Report ID 1317251514. I tried both standalone and FF plugin (with updated translators).

The entry I wanted to save is http://friaassociationer.blogspot.com/2012/01/metodologisk-pluralism-inom.html

My Zotero plugin version is 3.0b3.

beogl

Forgot: Zotero says "Save to Zotero (Blogger)". Note that the entry is on Blogspot, not Blogger.

adamsmith

Blogger is the name of the service, blogspot the address of the sites.
Apparently Zotero doesn't appreciate the international version here - we should be able to fix that pretty easily - I'll let you know when it's up.

beogl

I just tried again and it did not work. id = 341911452

beogl

I also tested this one now (I actually created the report id from that):

http://friaassociationer.blogspot.com/2012/01/metodologisk-pluralism-inom.html

adamsmith

as I say:

I'll let you know when it's up.

beogl

Oh, yes. Sorry ;-)

adamsmith

the fix is now up. Unfortunately it won't work on the links above, since those have moved to a non-blogspot URL, but generally this will work better for non English blogs now.

beogl

Thanks for notifying me, adamsmith.

Can't the meta tag here be used for recognizing the page type:

<meta content='blogger' name='generator'/>

adamsmith

the problem is that Zotero recognizes pages in two steps:
1. It looks for a pattern in the URL
If it recognizes the URL pattern it
2. looks for a patter in the page itself

If Zotero were to look for multiple patterns on every loaded site it would significantly slow down browsing. There are only 4 different, very widely used general patterns that Zotero looks for in all pages.

beogl

Oh yes, forgot. But a simple solution for that would be that it did not look for the pattern in cases like this until I want to save the page.

adamsmith

While theoretically possible, I don't see a good solution to actually make that work.
Either Zotero would have to run through the detectWeb functions of all 300+ translators when you say you want to save the page (and that actually wouldn't work, since some detect functions are quite generic (because the URL target regex is very specific) and will always return a value) or the user would have to select among those 300+ translators which seems like it'd create more chaos than benefit.

The last option would be to have a 2nd level detect function - i.e. a function that just looks for unambiguous patters like the one you have above - but that'd be an entirely new thing and I'm not convinced the pay-off would correspond to the outcome - how many sites apart from blogger would that actually apply to?

On a different note - maybe we _can_ come up with a useful and sufficiently restrictive target regex for blogspot posts on non blogspot domains -- looks like they're all /\d{4}/\d{2}/[\w-]+\.html

beogl

An easier way would perhaps be to mark some of the translators to run in the 2nd level detection function.

adamsmith

well - but that's a whole layer of functionality that doesn't currently exist in Zotero and would have to be coded - we're not just talking translators anymore. So that would have to have a significant added benefit to be worthwhile and as I say, I currently don't see that.