Simple Update Protocol (SUP)

Paul Buchheit (the creator and ex-lead developer of Gmail) of FriendFeed has suggested a new format – Simple Update Protocol (SUP) to minimize multiple feed polling.

The idea is simple -

• The feed provider will generate a live or cached feed for ALL recent updates with an ID field to identify the updated entities.
• The feed consumer will poll this SUP feed instead of polling individual RSS feed.
• The feed consumer will request only relevant updated feeds.

HTTP header If-Modified-Since does something similar. But, that would still means requesting for each individual feed.

The proposal looks good for both feed generators and consumers, and not very difficult to implement either. However, the format is too vulnerable to feature creep and standardization can be a painful process.

Lack of External Links in RSS Feeds

Full vs partial RSS feed is an old debate. I understand that a site needs to accumulate page views and ad impressions to pay for bandwidth. Now, Digg’s RSS feed and Slashdot’s RSS feed – both these sites don’t have much original content apart from the story comments. (Reading Slashdot comments is fun, even a few trolls. Digg comments OTOH are invariably retarded). These sites are not obliged to put all “their” content in RSS feeds, and may be that’s just me, but throwing an external link or two won’t hurt the click-through much and they can always put ads in feed.

RSS vs Atom Revisited

Atom is technically superior, more comprehensive and has more possibilities. From an engineer’s (and a purist’s) point of view – agreed. But RSS

• has been widely adopted already,
• has become synonymous to content syndication, and
• is more evolutionary than anything else out there.

I’m too naive to take a stand, but RSS just works.

Live Clipboard for the Web

Microsoft’s Ray Ozzie has come up with the idea of live clipboard for the web. I started watching screencasts skeptically thinking the idea to be another creation of the MS hype machine, but eventually I ended up watching all five of them.

The live clipboard idea looks appealing, implementable and probably the way to go. All it would need is that structured metadata is understood across websites, applications and operating systems. I guess that’s too much to ask, considering the inhibitions against everything Microsoft does with respect to standards. But the idea in itself is compelling, and it’ll be useful even if it can just be implemented across websites (watch #1). And we can expect at least Windows OS support if the idea gains some ground. Dave Winer has a mention of this too.

Standards these days are tricky things. We don’t even have a standard for syndication as yet.

Memetracker

A nifty new Web 2.0 buzzword. What’s a Memetracker? Let me define it first.

As communities grow in the online world new ideas emerge, spreading across the content fabric of the digital world at a sometimes furious pace. The speed of the emergence, and the following crash, of these ideas, or memes, makes tracking emergence by hand, or by an unaided individual, almost impossible. The sheer size and dynamic of the myriad communities makes tracking community information and ideas a data flow and logistics problem. Hence the need for memeTracker, a program to automatically crawl and analyze online communities, reporting the emergence and diffusion of new ideas.

The word ‘Memetracker‘ isn’t exactly new. The page I lifted the above text is resonably old, perhaps as old as 2002 – it isn’t even online now (Google cache link)

Robert Scoble asks ‘Why don’t you use Memetracker’. I do. tech.memeorandum is my favourite. TailRank being the other one. And while I do use memetrackers, I can understand why people won’t use it.

The ‘blog-oh-sphere’ is a place where there are more content creators than consumers, and everyone is busy honking their own horn – me included. The issues with memetrackers of today

1. There aren’t enough memes to sustain just blog based news ecosystem. So traditional sources of news are going to be part of memetracking, which narrows down its advantage over Google/Yahoo! News et al.
2. By definition ‘Meme’ are old news, and among the early adopters, almost everyone is so proactive to have accessed it earlier than ‘memtrackers’ pick it up.

Memetrackers are going to pick up, as aggregation is the key to content access in future. But so is customization. Memetrackers are going to drive and be driven by the growth of RSS, as those can provide customized remixing of news and related content – synergies of the two.

Some blurb from the link I produced earlier.

The Data Conduit

MemeTracker as a data service required the formulation of the program as a data conduit, a program able to be placed into an existing flow of data without significant alteration of the existing data paths or data expectations. This is facilitated by the existing XML database formats used by memeTracker. The data conduit model is detailed in the final diagram of the programatic flow of memeTracker in relation to other applications. The key to the data conduit model is the seperation of all processed data into individual chunks of information that can be consumed individually or as a whole data structure, without corrupting the integrity of the data as a whole.

But what do we have there to make financial sense of it – ads, not innovative enough.

Fixing RSS – II

Scott Karp is tackling two different extremes. On one hand there is ignorance about ubiquitous RSS feeds among most of Internet users, and then there is issue of information overflow with too many feeds for some. Fixing one of them may even worsen the other pain.

Dave Winer talks about ‘River of News‘ mode of feed aggregation. Which is a great idea to fix the feed abundance problem, but has its drawbacks. ‘River of News’ model seems to be too news-centric, which is okay if you are aggregating news and blogs. Syndication is not just about news.What about syndicating a wiki, or a novel, or may be research data or online learning material, which you won’t wish to miss in the ‘conveyor-belt sushi‘. The no-brainer is to separate the two – a river and a pond.

I’m yet to use Dave’s aggregator, but the idea is compelling. Google Desktop’s web clips come close to the idea – though Dave insists on a web interface, a desktop interface would be nice. But it’s what I think it is, I’d still need a pond-type aggregator.

Fixing ‘Really Simple Syndication’(RSS) for good

No one knows what 'syndication' means, unless you’re talking about I Love Lucy reruns. Syndication is a publisher-centric, geek-centric term. For most people, it’s Really Simple Huh? Most people don’t even know that syndicate can be used as a verb!

And then there are issues at the other extreme too – the problem of abundance of RSS feeds.

Scott Karp at Publishing 2.0 suggests this three step solution

1. Call it 'subscribing'

because 'subscription' is something most people are familiar with

2. Encourage everyone to get a reader

because most people either don’t have one or don’t know that they have one

3. Use the iTunes model — Search, browse, recommend, remix

an Amazon-esque 'people who subscribed to this also subscribed to…'

I would add,

4. One click subscription

The way we do it now is just too geeky. No wonder only 4% of Internet users know what RSS is.

So what will it take to 'feed' them RSS. Either,

1. Wait for Microsoft to fix this – for those who think that the little blue 'e' icon is THE Internet, or,
2. Make it easier than browsing itself, as easy as e-mail. (RSS integration in Outlook is soooo (a))

We'll see!

RSS Integration in Outlook 12

Michael Arrington at TechCrunch has posted that next Outlook release will have RSS integrated as announced by Outlook program manager Michael Affronti. Surely, a step towards widening the user base of RSS.

It’s easy to make people try out new technologies on a browser with hype and curiosity, but majority of such audience is ephemeral.

Bring it to the desktop – if it’s truly valuable, it’ll change people’s life.

Simple Sharing Extensions: Turning RSS bidirectional

Ray Ozzie, the Microsoft CTO, has proposed Simple Sharing Extensions(SSE)- a new specification that extends RSS from unidirectional to bidirectional information flows.

From the FAQ page on SSE at MSDN

SSE defines the minimum extensions necessary to enable loosely cooperating applications to use RSS as the basis for item sharing–that is, the bidirectional, asynchronous replication of new and changed items among two or more cross-subscribed feeds.

Essentially, an RSS (and OPML) extension to allow two-way communications over RSS*.

The draft specs for SSE are released under a Creative Commons license – Attribution-ShareAlike. By the way, Creative Commons has made a note of SSE and its CC-license release, saying something to the effect of it’s too good to be true.

For me, SSE is a potential blockbuster. Too early to comment though.

*Is that Really Simple Sharing in OzzieSpeak or Really Simple Syndication?

Google Reader

Now this one was due from Google, and appeared without the usual Google Chaos. A RSS Feed Reader competing with Bloglines, My Yahoo!

Being web browser based it has all the usual drawbacks and advantages. AJAX goodness on one hand and “Arrgh! The page has been corrupted” on the other.

Not a completely polished product yet, e.g. very few loading messages to notify user of any processing as in Gmail ( Alex Bosworth‘s Ajax Mistakes #1)

OPML importing from current feed reader(Suace Reader) failed on the first time with adding just one feed, and the second time it took ages to import. I am still not sure it has imported everything.

As of now it doesn’t seem usable enough for me. Even Google Sidebar is much more useful than Google Reader.