how does vienna tell if a rss item is differnent?

An RSS/Atom newsreader with features comparable to commercial newsreaders.
andreiv
Harmless
Posts: 1
Joined: Wed Mar 19, 2014 1:23 pm

how does vienna tell if a rss item is differnent?

Postby andreiv » Wed Mar 19, 2014 1:28 pm

Hi, I have a vienna user who mentioned that one of our feeds keep feeding duplicate items.

Our stories get edited a lot, could that be an issue?

How does vienna decide if items are different to show it again in the reader?

Does it diff on item text/content as well?

Thanks

barijaona
Vienna Team
Posts: 626
Joined: Sat Nov 12, 2011 11:10 am
Contact:

Re: how does vienna tell if a rss item is differnent?

Postby barijaona » Thu Mar 20, 2014 4:44 am

Hi,
First, Vienna tests the guid element if there is one.
If no guid is available, Vienna checks the title and the date/time of the post.
I contribute to Vienna RSS as a developer.
Please, don't forget those tips for writing a good bug report

Deimos
Harmless
Posts: 15
Joined: Sat Mar 29, 2014 10:51 am

Re: how does vienna tell if a rss item is differnent?

Postby Deimos » Sat Mar 29, 2014 10:09 pm

barijaona wrote:Hi,
First, Vienna tests the guid element if there is one.
If no guid is available, Vienna checks the title and the date/time of the post.


Does Vienna use the <updated> tag in Atom xml feeds ?

barijaona
Vienna Team
Posts: 626
Joined: Sat Nov 12, 2011 11:10 am
Contact:

Re: how does vienna tell if a rss item is differnent?

Postby barijaona » Sun Mar 30, 2014 1:38 am

Deimos wrote:Does Vienna use the <updated> tag in Atom xml feeds ?

Yes
I contribute to Vienna RSS as a developer.
Please, don't forget those tips for writing a good bug report

Deimos
Harmless
Posts: 15
Joined: Sat Mar 29, 2014 10:51 am

Re: how does vienna tell if a rss item is differnent?

Postby Deimos » Sun Mar 30, 2014 9:44 am

barijaona wrote:
Deimos wrote:Does Vienna use the <updated> tag in Atom xml feeds ?

Yes


I asked because I have my doubts and looking into it I now suspect it is not working properly. I have been suspicions on a number of feeds (from the same site) where I used to get loads of updated articles flagged but since changing to Vienna have never seen a single updated article on the site. I have about 10 feeds from the site (all behave the same) but a good example is https://www.lonelyplanet.com/thorntree/forums/americas-central-america.atom. Just checked and in Vienna there are no updates shown yet the XML shows an updated article originally posted and read yesterday updated a few moments ago (according to the <updated> tag), yet Vienna shows it as an old article from yesterday (with no update indicator). In fact loads of such articles when I just checked.

Same behaviour on all the other feeds from the site.

For me this is pretty urgent as I am missing loads on important posts.

Many thanks

barijaona
Vienna Team
Posts: 626
Joined: Sat Nov 12, 2011 11:10 am
Contact:

Re: how does vienna tell if a rss item is differnent?

Postby barijaona » Sun Mar 30, 2014 4:54 pm

Deimos wrote:I asked because I have my doubts and looking into it I now suspect it is not working properly. I have been suspicions on a number of feeds (from the same site) where I used to get loads of updated articles flagged but since changing to Vienna have never seen a single updated article on the site.


You need to enable the « mark updated articles as new » preference.
When this is enabled, updated articles are marked with a green dot in Horizontal and Vertical layouts. New articles are marked with a blue dot as usual.
I contribute to Vienna RSS as a developer.
Please, don't forget those tips for writing a good bug report

Deimos
Harmless
Posts: 15
Joined: Sat Mar 29, 2014 10:51 am

Re: how does vienna tell if a rss item is differnent?

Postby Deimos » Sun Mar 30, 2014 6:37 pm

barijaona wrote:
Deimos wrote:I asked because I have my doubts and looking into it I now suspect it is not working properly. I have been suspicions on a number of feeds (from the same site) where I used to get loads of updated articles flagged but since changing to Vienna have never seen a single updated article on the site.


You need to enable the « mark updated articles as new » preference.
When this is enabled, updated articles are marked with a green dot in Horizontal and Vertical layouts. New articles are marked with a blue dot as usual.


Did that when 1st installed. Getting green "updated" flags on some other feeds, but none on the site in the example given (and all other feeds from the same site). The site given has quite a lot of updating going on (confirmed when you look in the XML file).

The difference between feeds that show updated indicators and none seems that the RSS feeds show the updated indicator whereas the atom feeds don't. Not checked every feed I use (as some have infrequent updates and you need to check XML straight away, etc.)

barijaona
Vienna Team
Posts: 626
Joined: Sat Nov 12, 2011 11:10 am
Contact:

Re: how does vienna tell if a rss item is differnent?

Postby barijaona » Mon Mar 31, 2014 4:01 am

Thanks for reporting this. I will investigate.
I contribute to Vienna RSS as a developer.
Please, don't forget those tips for writing a good bug report

Deimos
Harmless
Posts: 15
Joined: Sat Mar 29, 2014 10:51 am

Re: how does vienna tell if a rss item is differnent?

Postby Deimos » Wed Apr 02, 2014 1:33 pm

FYI: The site developer (who I have also been discussing this with) believes there may be issues with the <updated> tag in the header section (i.e. outside the <entry> sections. They are checking and will let me know if/when site changes have been implemented. They feel that if the header <updated> is sometimes suggesting the XML has not changed then an RSS Reader might not bother to look at the individual <entry> sections.

Just to update you in case it impacts your own checks.

Deimos
Harmless
Posts: 15
Joined: Sat Mar 29, 2014 10:51 am

Re: how does vienna tell if a rss item is differnent?

Postby Deimos » Fri Apr 04, 2014 8:53 am

Developer have changed the header date and Vienna is still not showing articles that are updated.

A good example is https://www.lonelyplanet.com/thorntree/forums/americas-central-america.atom. Which is quite an active feed, lots of articles being updated but a reasonable number of new articles so is is manageable to investigate (without ploughing through vast volumes of data.

It looks like only new articles are being shown (i.e. the <pyblished> tag date being used) and the <updated> tag date being ignored.

Truerob
Harmless
Posts: 1
Joined: Fri Apr 04, 2014 9:24 am

Re: how does vienna tell if a rss item is differnent?

Postby Truerob » Fri Apr 04, 2014 9:32 am

barijaona wrote:Hi,
First, Vienna tests the guid element if there is one.
If no guid is available, Vienna checks the title and the date/time of the post.


Hi,

ATOM uses <id> tag to identify resource, <guid> is for RSS (probably not an issue, just making sure we are good on this one).

The id/guid doesn't change (and shouldn't since the resource is still the same), just the <updated> tag. Could this be the case?

thanks,
Rob

barijaona
Vienna Team
Posts: 626
Joined: Sat Nov 12, 2011 11:10 am
Contact:

Re: how does vienna tell if a rss item is differnent?

Postby barijaona » Mon May 05, 2014 12:21 am

Hi Deimos and Rob,

Sorry for taking so much time to look at this.
In fact, Vienna checks the article title and article body to make sure that they have changed ; if they have, the article is marked as updated.
This is reasonable for most feeds (blogs and forums where each contribution is an article), but this does not fit will with these Lonelyplanet forums : the article body does not include the part quoting « Last reply was … by … », only the original question.
So Vienna does not see there was a change.

After reading the Atom specifications (RFC4287), I am just wondering if I should change this… I am afraid many feeds routinely change the « updated » information, and trusting it might create new problems…
I contribute to Vienna RSS as a developer.
Please, don't forget those tips for writing a good bug report

Deimos
Harmless
Posts: 15
Joined: Sat Mar 29, 2014 10:51 am

Re: how does vienna tell if a rss item is differnent?

Postby Deimos » Tue May 06, 2014 1:50 pm

barijaona wrote:Hi Deimos and Rob,

Sorry for taking so much time to look at this.
In fact, Vienna checks the article title and article body to make sure that they have changed ; if they have, the article is marked as updated.
This is reasonable for most feeds (blogs and forums where each contribution is an article), but this does not fit will with these Lonelyplanet forums : the article body does not include the part quoting « Last reply was … by … », only the original question.
So Vienna does not see there was a change.

After reading the Atom specifications (RFC4287), I am just wondering if I should change this… I am afraid many feeds routinely change the « updated » information, and trusting it might create new problems…


Many thanks for investigating and posting what you found.

From my Windows development days I would say "Make it a settings option". But since moving to OS X, I appreciate that Apple Inc's method is to remove options forcing everybody to work in the same way as the developer (something I don't agree with myself).

I've not read the Atom spec, but having examined the XML, I would say (in priority order):
1. Check and process the <updated> date for each article (which may cause issues for some feeds - but they should set the <updated> date correctly.
2. Make the "Mark updated articles as new" preference a feed setting rather than application wide, so the behaviour can be adjusted on a per feed basis. Maybe, if some feeds are "ill-behaved", introduce a per feed setting "Base article change on Title/Body" or "Detect article update usning Atom <update> tag date" - but phrased so others would understand.

I would say the 1st change is the most important (at least that way properly behaved feeds are processed properly rather than correctly behaved feeds not working whilst ill-behaved "work"). No urgency to include both changes in the same release.

Those are my initial thoughts.

Anyway, thanks for investigating and responding.

Deimos
Harmless
Posts: 15
Joined: Sat Mar 29, 2014 10:51 am

Re: how does vienna tell if a rss item is differnent?

Postby Deimos » Wed May 07, 2014 12:06 pm

Deimos wrote:
barijaona wrote:Hi Deimos and Rob,

Sorry for taking so much time to look at this.
In fact, Vienna checks the article title and article body to make sure that they have changed ; if they have, the article is marked as updated.
This is reasonable for most feeds (blogs and forums where each contribution is an article), but this does not fit will with these Lonelyplanet forums : the article body does not include the part quoting « Last reply was … by … », only the original question.
So Vienna does not see there was a change.

After reading the Atom specifications (RFC4287), I am just wondering if I should change this… I am afraid many feeds routinely change the « updated » information, and trusting it might create new problems…


Many thanks for investigating and posting what you found.

From my Windows development days I would say "Make it a settings option". But since moving to OS X, I appreciate that Apple Inc's method is to remove options forcing everybody to work in the same way as the developer (something I don't agree with myself).

I've not read the Atom spec, but having examined the XML, I would say (in priority order):
1. Check and process the <updated> date for each article (which may cause issues for some feeds - but they should set the <updated> date correctly.
2. Make the "Mark updated articles as new" preference a feed setting rather than application wide, so the behaviour can be adjusted on a per feed basis. Maybe, if some feeds are "ill-behaved", introduce a per feed setting "Base article change on Title/Body" or "Detect article update usning Atom <update> tag date" - but phrased so others would understand.

I would say the 1st change is the most important (at least that way properly behaved feeds are processed properly rather than correctly behaved feeds not working whilst ill-behaved "work"). No urgency to include both changes in the same release.

Those are my initial thoughts.

Anyway, thanks for investigating and responding.

And, having thought a bit more about it; I would say that you could just add the check to the Atom <updated> date to the existing title and body checks. Any change to either the title or the body indicates an update, as does the <updated> date time. should any f those change then it would seem likely that the article has changed.

So maybe ignore all my Preference settings, etc. and just add an additional check on the <updated> date.


Return to “Vienna”

Who is online

Users browsing this forum: No registered users