awasu.user wrote:Well, when I start FeedDeamon I watch how many new all item is to calculate how many time I will be work. I have feed grouped by topic in folder. It's more useful for me info how many new articles I had in folder than unread feeds.
No worries. I'll take a look at how much work it would be to implement this.
awasu.user wrote:When I add feeds from file each channel was update time for one day as default. When I change this in main Options it was nothing to happened.
Yes, this only applies to new channels.
awasu.user wrote:I was lazy and I reset all data by Awasu Admin, then I want change default update time and next import file. I make mistake and I delete user account and I can't run Awasu. Could you tell me how can I change my mistake?
I have info: "No user account were found! If you are sure Awasu has been used on this computer, please add the users manually."
Yes, if you run admin.exe it won't find your user, because you deleted it But if you run awasu.exe, it should start up with a clean slate. You'll be asked to enter your trial key, you can use the same one as before.
awasu.user wrote:I need run Awasu twice time a day, morning and afternoon. When Awasu is running I want auto update after 15 minutes until I not choose all link to proceed before first selection and on second I send choosen to Instapaper from browser.
Keep an eye on the Activity Log, check the timestamps, I think what Awasu is doing is what you want.
awasu.user wrote:Could you explain me how "update all feeds" work?
I think when I push it it will update now, but in my test it's looks like first wait some time and after that doing something.
Yes, it will update all channels immediately. Watch the Activity Log and you will see Awasu starting to update all the channels.
awasu.user wrote:Options for channel - frequency of updates - how determine this? It is a delay before update or after update of feeds?
If a channel is configured to update every 15 minutes, and it updates at 09:00, the next automatic update will be at 09:15. If you do a manual update at 09:05, the next automatic update will be at 09:20. If you close Awasu at 09:10 and restart it at 13:00, the 09:20 update will be overdue and the channel will update as soon as Awasu starts.
awasu.user wrote:I can't fully understand your troubles. I keep in mind not complicated calculation of similarity of text based on frequency used of words or something that.
Why is uniquely identifying items difficult? Ah, let me count the ways...
- Even if 2 items have identical text, there's extra metadata that's not always visible that's part of the item. If the "# views" are different, that's maybe not enough for them to be different, but what about "author"? Or "published time"?
- What happens if a publisher changes an item after it has been received by Awasu? Does adding an extra space to the text trigger a full recalculation of dupes? What about metadata?
- Some publishers put out feeds where every URL is the same. And if you think that's just silly and their fault, there are cases where it's not possible to have a sensible URL e.g. what if the channel is a plugin retrieving data from a database or email account - what should the URL for each item be? These items might have a different GUID, but then Awasu would consider these items to be always different (since a GUID is supposed to be unique ID), so deduping wouldn't work if there were 2 items that actually were the same.
- What if 2 items are identical except for their GUID's - are they the same? Most people would say they are the same, but the specs say they are different. What if one has a GUID but the other doesn't?
- Likewise, what if 2 items have the same GUID, but different content? The spec says they are the same, regardless of the content.
- If 2 or more items are considered to be the same, which one do we show and which do we hide?
- If we have a set of items that are considered to be the same and have hidden all but one, what happens if one of the hidden items changes? Does it reappear?
- If Awasu has hundreds of thousands of items, or more, what is the performance impact of deduping?
I'm sure there's more Yes, a lot of these are corner cases, but there are a lot of low-quality feeds out there because everyone thinks it's easy and so roll their own code to generate them (and the RSS specs are not particularly good, but that's another rant for another day). Being a programmer is so much fun...
TBH, I don't think I've ever seen cases where exactly the same item is published in different feeds. Yes, people reblog stories all the time, but there's always something different e.g. the URL, or a "reblogged from..." comment at the end, something that would prevent them from being detected as dupes.
awasu.user wrote:Second step for me is check than feed item NAME (article title) is the same and it is all. For example item feed from NYT is about change taxes and we have it on magazine feed and Business feed. I wrote about the same text with identical title, description, url and remove it from item list after push the button.
If you just want a simple check of title/URL, it would be possible to write something that queried the API to get a stream of incoming items, and dedupe using this simpler rule.
If you want hide items from a channel, you could configure it to show only unread items, then mark items as read when you want to "remove" them.
BTW, the workflow people often use for this kind of thing is to use workpads. They are somewhere you can put items for working on later, then generate reports from, etc.