Super duper tl;dr:
- Christensen's theories of industry disruption fit digital publishing like a glove 
- Marco's The Magazine is a pretty good 'subcompact' example 
This essay is the written + expanded version of a talk I gave at The Internet Archive (Books in Browsers) last month:
Watching the video or reading the essay both take about the same amount of time, but the essay has much more information.
You can skip to my UI/UX deconstruction of The Magazine here:
Or, skip to the 'Subcompact Manifesto' here:
 Innovator's Dilema; Also: this Neiman Report: http://www.nieman.harvard.edu/reports/article/102798/Breakin...
Online publishing nonsensically adopts many of the practices of offline publishing.
It doesn't need to, and could benefit from adopting "Small issue sizes (3-7 articles / issue), Small file sizes, Digital-aware subscription prices, Fluid publishing schedule[s], Scroll [instead of pagination], Clear navigation, HTML(ish) [formatting], [and] Touching the open web [not just platforms]".
Issues is one of those throwbacks to physical publishing. When you have to physically print and distribute, issues are a good thing... but online you can just publish an article when it's ready.
You mentioned how you think "edge" is important, so I would guess one of your arguments might be that a "publish on completion" stream is the exact opposite of edge. However, couldn't you just create artificial edges? For instance, showing the articles for the past 7 days?
Mainly though, they limit the number of times a reader is pinged. I'm a big advocate for grouping all notifications, not just 'content' — only notify me if #_notifications > x. Issues can be a kind of inherent frequency limiter.
That said, they're absolutely not necessary and most online publishing (rightly) has no concept of issues.
"However, couldn't you just create artificial edges? For instance, showing the articles for the past 7 days?"
Yes — and I think more apps should allow for this.
Facebook is a great example (IMO, but clearly not in the opinion of many others) of doing just this — taking a near infinite stream of one-offs and grouping them based on your Facebook use frequency / habits. Twitter's (Summify) summary articles are another good implementation of this: what happened in my stream yesterday?
Second (though related), what is the place of advertising in this theoretical subcompact publication?
In the print world, you have to charge each reader because there's an actual cost associated with producing each physical copy, but this is not the case for digital products, where distribution is effectively free. And since free crowds out paid, it's extremely difficult for a paid digital magazine to build up a following. (Thought experiment: if this article itself had been published behind a paywall, how many would have read it, and would it have ended up on HN?)
Now, of course there is still a cost associated with actually creating quality content. The key to success will be finding a way to pay for it without charging readers and thus kneecapping distribution.
Too short, read more: http://gyrovague.com/2012/06/28/why-the-web-will-gut-paid-e-...