Actually emski and Markus are wrong.
Well, I do not actually know if they are wrong, but there are smartasses that say that we cannot continue like that! And we must set a bar for the quality of the price feeds. If the price feed is below the said price feed standard (read price) it will no longer be valid (cause they said so!)
The argument goes something like this -"Do you know what will happen if the quality of price feeds go below 90% quality? They will gradually continue to decline until the reach 0, quality...", "Nobody will trust the feeds and they will eventually reach zero".
That's why we will set the standard... and no more crazy non-standard feeds will be allowed.
And it will be awesome after that! <hearts><hearts> many <hearts>
PS
And Dude, if you do not trust the quality feed theory you should go start your own blockchain, experimenting with such unsafe propositions... we are quality feed believers here (well we became such recently, but now in quality feeds we trust!)!
I'm genuinely interested in your definition of price feed quality.
What features is it supposed to have?
My idea of a perfect price feed is that:
Get all "real" trades. Average out weighting in the volume of each trade. The result is "perfect" feed.
As this is impossible due to many reasons I think a solution that takes into account the volume as well as each price of the exchanges is as good as it gets.
There are some other issues with price feeds:
1 If an exchange fails to provide its feed data and this influences significantly the feed => should we still publish the new feed or wait for the exchange to resume its feed service? (my version of the script already allows configuration of this behavior)
2 If an exchange publishes significantly higher/lower price for an asset at tiny volume => should we account that?
Currently if you want to verify my feeds you can just run the script yourself (my settings are in the repo). If you believe my script (or anyone else's) isn't accurate and should not be used => state it and explain yourself.
On the other matter you pointed out: Should we use price feeds at all? It would be a good idea if someone
briefly describes all the possible alternative approaches and maybe start a discussion?