"An excellent book could be written about this period, about how Dave changed the internet."
There may have been some interesting ideas coming out of his shop, but isn't this a little ridiculous? I don't think a single one of these products ever got any real traction, and, frankly, most of them were flops.
I was using Macs during this period. I tried the thing I think was called Radio: started it, it insisted on getting an email address, so I typed mine in. Not more than 10 minutes later I get a spam email from some clown who was handed my email address by Radio, which apparently broadcast it to other Radio users all over the world without asking permission. I deleted it from my machine and never looked at another piece of Winerware after that. But I kept up with comments about Userland stuff on the internet, and noticed a barrage of complaints about security, privacy, memory leaks, instability, and the infamous, irrational behavior of the CEO.
RSS was invented at Netscape with no involvement by Winer. Most informed people starting fresh today would tend to prefer the Atom format.
"XML-RPC, SOAP, OPML"
Some of the "interesting ideas" that I acknowledged. Most people today would prefer REST if possible, although there is a lot of legacy and enterprise stuff that uses, especially, SOAP. OPML is obviously useful, if obvious.
"arguably podcasts were first implemented by Winer"
Automatically downloadable audio files were already in place in several forms before Winer got involved, including subscriptions. Winer and Curry implemented RSS enclosures and used them to deliver audio files. This is an implementation of the existing podcast concept using RSS, and probably helped make podcasting more popular.
The RSS spec is something that a motivated high school student could come up with in a couple of hours. Why grown men consider it an accomplishment and argue about who should get the credit is embarrassing. The real value comes in building a consensus to create an accepted standard. Did Winer do more to help this process along, or to obstruct it? What about stunts like making trivial changes to an existing spec and slapping Userland's copyright on it? What effect did that have on helping to establish the RSS spec as a standard? Edit: The Atom format owes its existence mainly to the desire to create a work-alike that avoids the RSS version mess, which resulted from Winer's antics. Mark Pilgrim early on defined Atom as "RSS without the sociopaths."
And I referred to "products" as flops; you mentioned standards and protocols. The products I'm referring to are Radio and its many precursors and descendants. Do you think that any of these products were successful?
The thing I've never been able to get past with Winer's tech is that he is so married to hierarchical representation. I've always felt like my own thoughts are best expressed as graphs.
edit: I mean CS graphs; many-to-many; a child having multiple parents. I don't mean your basic mind-mapping graphical software, which very often also don't support representing graph data structures.
Many outliners support graphs... look for the "clone" feature, though it's called various things in various outliners. The problem is that you have the same sort of problem you do with UNIX filesystems, where almost everything expects to work on trees, and then, suddenly, bam it's not a tree anymore and something a little funky or unexpected happens because some code expecting a tree is suddenly being run on a full graph.
From "UNIX Implemenation", _The Bell System Technical Journal_ Vol 57, No 6, Part 2 (July-August 1978):
"The file system structure allows an arbitrary, directed graph of directories with regular files linked in at arbitrary places in this graph. In fact, very early UNIX systems used such a structure. Administration of such a structure became so chaotic that later systems were restricted to a directory tree."
No "sort of" and no graph-"like", it's just, yes. If you follow symlinks you can have a graph. Many filesystems also permit hard-links to form a graph, which makes it that much trickier for a program to determine if it's in a graph or tree regime; I believe many file systems also deny this.
Symlinks + preventing directory hardlinks are a decent compromise in power, where things that choose to understand them can handle graph-like structures, but things that wish to treat filesystems naively (thus treating symlinks as just another file) can work on trees. Beware the "friendly" library that automatically follows symlinks in something like a scripting language, for they are throwing you to the wolves in the name of user friendliness.
Been a while since I thought about Scripting.com. It was the first "blog" I really read. I thought I was so cutting edge: had a reader app on my Palm (Sony Clie) that I synced and could read offline. Good times.
There may have been some interesting ideas coming out of his shop, but isn't this a little ridiculous? I don't think a single one of these products ever got any real traction, and, frankly, most of them were flops.
I was using Macs during this period. I tried the thing I think was called Radio: started it, it insisted on getting an email address, so I typed mine in. Not more than 10 minutes later I get a spam email from some clown who was handed my email address by Radio, which apparently broadcast it to other Radio users all over the world without asking permission. I deleted it from my machine and never looked at another piece of Winerware after that. But I kept up with comments about Userland stuff on the internet, and noticed a barrage of complaints about security, privacy, memory leaks, instability, and the infamous, irrational behavior of the CEO.