ubuntu-touch-coreapps team mailing list archive

[Christian Dywan <christian.dywan@xxxxxxxxxxxxx>]

I might suggest to use two separate databases. One for subscribed feeds
and read status, synched. Another that stores feed contents as a local
cache only.

ciao,
Christian

-----Original Message-----
From: Roberto Alsina <roberto.alsina@xxxxxxxxxxxxx>
To: David Planella <david.planella@xxxxxxxxxx>
Cc: Lisette Slegers <lisette.slegers@xxxxxxxxxxxxx>,
ubuntu-touch-coreapps@xxxxxxxxxxxxxxxxxxx, ubuntu-phone
<ubuntu-phone@xxxxxxxxxxxxxxxxxxx>, Christian Dywan
<christian.dywan@xxxxxxxxxxxxx>
Sent: Fr., 12 Jul 2013 16:35
Subject: Re: [Ubuntu-touch-coreapps] [RSS reader] Saving offline

On Fri, Jul 12, 2013 at 10:08 AM, David Planella
<david.planella@xxxxxxxxxx <mailto:david.planella@xxxxxxxxxx>> wrote:

    On Fri, Jul 5, 2013 at 5:18 PM, Lisette Slegers
    <lisette.slegers@xxxxxxxxxxxxx
    <mailto:lisette.slegers@xxxxxxxxxxxxx>> wrote:

        Hi everyone!

        During the RSS hangout today, we talked a little bit about
        saving offline, and there are some interesting questions we
        should consider. 

        If the app currently saves all the articles in the reader's
        stream into the local database, how is the app affected if:

         1. The reader subscribes to *a lot* of feeds.
         2. The reader has used up a lot of storage on his phone
            (storing music and videos for example).

        Could the user run out of space in the scenarios above? What
        would happen in that case; would you prioritise new articles and
        chuck out old ones? I guess that would be fine as long as your
        history has an acceptable size, although it might be confusing
        for a user if the app sometimes saves 2 days of history, and
        sometimes 2 hours if there is not enough space on the device. 

        Imagine the reader with a lot of subscriptions. His feed stream
        is more like a tsunami. He skims the article headlines very
        quickly and only opens a small percentage to read the first
        paragraph. Even fewer articles he saves to read in full at a
        later stage. Perhaps headlines are more important than the
        entire article? As long as a user has not saved an article, I
        think it is fair to assume we have an internet connection
        available to load the article. Or perhaps it could be a mix of
        online and offline depending on the amount of available storage.

        It would also be good to know whether we can technically save
        articles offline including the images. If it is not possible for
        version 1, it would still be useful to let users bookmark an
        article so that they can easily find it back later. As long as
        we use correct wording to clarify functionality to users
        ('bookmark' vs. 'save offline').


    CC'ing Roberto Alsina from the U1 team and Christian Dywan from the
    SDK team, who's the expert on the U1DB QML bindings.

    @Roberto, Christian, do you think that could be a good use case for
    U1DB storage? Do you think it might be a good idea to store RSS
    articles (text + images) in U1DB for bookmarking them? Or do you
    think offline reading support should be only on a per-device basis
    and better stored in a cache?


I would use U1DB for storing things like "favourites" and what articles
have been read or not, but not for offline 
reading, because the cost of syncing is probably offset by the
inefficiency, since articles are mostly read once, in one device, right?

So, syncing that same article to N devices more or less guarantees N-1
copies are never going to be read.

Attachment: signature.asc
Description: OpenPGP digital signature