For several days I have been trying to get a clear, concise response to what I thought was a very simple question. Apparently other Newsblur clients also believe it’s a valid question, based on their posts to the two threads I have posted it in.
Samuel, I have come to the conclusion that you either:
Don’t want to answer the question, because you know the answer is not what your paid clients want to hear.
You don’t understand the question, in which case the burden falls on me to simplify the way it is worded.
Can you help me understand which of those might apply so that we can settle the issue?
I’m tired of asking the question, and I imagine that others are tired of seeing me badger you.
So, again, I’m asking: Is the following commitment made by you in March 2013 still valid?
"The scaling that I need to do to the backend is about a week away from shipping. It’s a very involved process that requires me to do a lot of prep work and testing. Once that’s in place, I will crank up the feed fetching schedule to what it once was, which is every feed held by a premium user is updated at the minimum of once an hour. Popular feeds more often, and if the feed has published less than an average of 1 story over the past month, then it gets updated just a few times a day. "
I’ve lost track of what your original yes/no question is, but I think it was something to do with the feed times, which as you’ve had explained to you isn’t a simple same-for-every feed process.
Since you’ve consistently refused to actually tell anyone which feed you’re having trouble with, how can you expect anyone to actually look into it and give you a straight answer?
If, after doing an instant fetch, you are still having issues with a feed not updating as often as it should, post the URL that appears in the address bar when you click into that feed. He needs specific examples of broken feeds to be able to help…
Given that he noted a feed was updating every half hour last night after his fix, I’d say the answer is probably “Yes”. Insta-fetch your feeds that still have issues and see if you’re still beating a dead horse or not.
I am not currently claiming that there is a fault. I am, however, trying to determine what the broader, overall policy is. Samuel posted a pledge, I am simply trying to find out if the pledge is still valid.
I have read through every update you have made on the matter.
While I agree you have been pretty repetitive, I also see that no one from Newsblur.com (presumably Sam) has taken the time to answer your query head on.
I also would like to hear the answer to your question.
You have posted a direct, succinct, and relevant question that is plainly a yes or no question.
In looking at my own feeds, I see the same concern that was raised in another posting, so it does seem to me that it is still a valid issue.
I will post my own feed data in the other message line.
How long is a piece of string? A simple yes/no answer will suffice. Not all questions can be answered with a simple yes or no.
You ask if “the following commitment” is still valid. What commitment? All I see is him saying what he intends to do, nowhere does he say every single feed will be like that, or that it’s guaranteed or committed to or a pledge or any of the other words you seem to like putting in his mouth.
It’s a single paragraph talking about a very complex system, it’s never going to be anything more than a grossly oversimplified generalisation.
I don’t know if it’s currently checking that often, I suspect it’s not there yet. If you want to know why a specific feed is doing what it’s doing, post more specifics. If you want to know the exact formulas as they run today, read the code. Beyond that, I have no idea what you want.
Maybe “Brick”, like me, doesn’t know how to “read code”? Earlier you referred to Github. I had to Google the word before I even knew what you were banging on about.
I would think his advice to stop reading the thread was pretty much on the mark, David.
I’ve come to the conclusion that Samuel will never answer the question directly, so I am therefore wasting my time (and others, it seems) by continuing the battle.
I can only conclude that the answer to my question is “No”. I base that conclusion on the fact that if the answer were “Yes”, he would have wasted no time in saying so.
A yes answer would have meant that Newsblur was:
Checking every feed held by a premium user once per hour
Checking popular feeds more often than once per hour
Checking sporadic feeds at least twice a day
If that were the case, he would have taken credit for living up to the expectation he set.
Since he has not answered “yes”, the only conceivable alternative is that his answer is “no”.
Now that I have deduced the answer, I have no reason to continue the conversation.
Unless, of course, he wants to tell me that I am wrong.
I have a statistics bubble that says “If you went premium, this site would update every 72 to 90 hours…”. That is a pretty good indicator that #3 isn’t taking place.
Also, as David pointed out, even scheduling of feeds is a complicated system. It would be easier for you to look at the code, written in Python and can be read nearly as plain English, than have me explain all the differences that bump a feed refresh interval up or down.
I’m thinking about introducing a statistics that shows a feed-fetch-lag. Effectively an average measurement of the difference between the fetch and the earliest unseen story’s date. Problem is that dates are specific to feeds, so the skew would be different per-feed. In other words, if the time was a 5m lag between post and fetch, but their server was misconfigured by 1h, it would always show as a 65m lag. Not sure how I can best capture this, but I’ll noodle on it.
This is a long-winded way of saying that I try to minimize that lag on the overall system. But there are always going to be feeds that publish once a week and NewsBlur will be 6 hours late to it. I’m boosting feed fetching regularly. If you notice it’s hit 6MM fetches a day, and I expect it to climb in a few weeks.