Microsoft is building a Google cluster

From Greg Linden's blog, a quote from Microsoft:

"the people who could build a viable [Web] services infrastructure of scale are companies that have both the will and the capacity to invest staggering amounts of money."

Hmm. I suppose that would exclude Google, BitTorrent and the Web itself. Unless they meant "the people that need to catch up will need staggering amounts of money, otherwise they lose".

dojo.storage: Client-Side Storage

From Ajaxian, here is a post about
offline access and client-side storage via Dojo. We are getting close to the tipping point for disconnected use and client-side state - and with the strength of the REST buzz (I can't believe how many people are actually applying REST!) this really bodes well for the next few years of Internet-scale application development. If nothing else, at least it's yet another way to route around the damage that is the Win32 API.

Pubsub .vs. polling - again

This is a great read from Bob Wyman on Dave Winer's comments about polling compared to notification :
From As I May Think...: Dave Winer: Show me that mathematical proof!

Dave's lack of understanding of the issues related to scaling can be seen in the history of the weblogs.com site that he struggled to build and maintain for so long. That site takes "pings" from blogs and then consolidates them into tremendous "change lists" which must be polled. Essentially, this site converts an efficient push-based update notification system (pinging) into an inefficient polling based system. Weblogs.com, as Dave built it, didn't even support common methods like eTags or RFC3229+feed to improve polling efficiency and scaling. The result was that it simply didn't scale and was frequently incapable of providing the service levels that people expected. Only now that Verisign has taken over the site and dedicated much better engineering staff and much more hardware, has the weblogs.com service begun to be somewhat useful again. However, since weblogs.com is still based on the terribly inefficient polling of change-lists that Dave supports, it is still a far cry from being what it might be.

Ouch. But I did like the nod to KnowNow and mod-pubsub. Not sure if mod-pubsub is still alive and kicking though.

Best Google Image

The latest home page graphic from Google is my absolute favorite -

Delta feeds - RFC3229+feed

About a year and a half ago, Bob Wyman was instrumental in defining an approach to greatly reduce the load and bandwidth used by applications that polled for changes to RSS/Atom feeds.
The other week, he noted that Microsoft will support the RFC3229+feed approach as well - which is good.

The only problem I have with this approach is that I think it is simpler to use hyperlinks, and I haven't seen a real comparison between the two. Both approaches have the client application maintain state of what data was last retrieved, but using hyperlinks has more chances for pre-existing caching servers to work without modification. I think the Atom protocol has defined something like this, but I couldn't follow the email threads.

To use hyperlinks, the data returned in a feed would have a link to the 'next' (more recently changed) posts. The client would then follow that link, which would either be empty (and optionally have a cache-control header to indicate how long to wait before checking again) or have more data - along with another 'next' link. The client just keeps following the links. The client would have two URIs - the original, well-known location that new readers start from, and the changing one which is the set of data most recently retrieved by that particular client. The server decides what the 'next' link is and what it contains - the data would be very cachable across all clients, merely by checking the URI.
The downside of this approach is the need to put the link within the content of the response - or add a response header for that location if the content isn't easily extended.

I put together a sample application that shows how this works - this is a simple html/javascript chat client, this is a link to the list of messages.

Broken as Designed

Sith Obasanjo in Broken as Designed can't decide which way is away from the dark side -

"However I do think some Web/REST advocates need to look around and realize what's happening on the Web instead of arguing from an 'ideal' or 'theoretical' perspective."

The Web advocates need to realize what's happening on the Web??

Okay, we all know some resources have broken content-type headers when you retrieve them. Others don't. Some clients never use content-type correctly, others do and many use heuristics. That's okay. Just do your homework - and part of that is to read the TAG finding and use it where appropriate - and be a good web citizen. That's simply practical advice based on practice. The "don't use content-type" is a theory that is unproven in practice.

Update - it looks like Dare found breakage in Cookies as well, but after further investigation it turned out to be a problem with the data returned by the server. But if the bloggers using RSS can't correctly control this Cookie response header - would we have no choice but to drop support for that header as well, as suggested for the Content-Type header? I mean, the theory of Cookies is all well and good, but if you look around to what the major players like MSN are actually doing, who are we to stick with something broken as designed merely because of theory?

Persistent Search and OpenSearch

It looks likes there's more interest (again) on saved searches and search alerts -
unto.net - Persistent Search and OpenSearch

Hopefully things have changed since I last reviewed this space in 2004.

Content-Type is dead

What a simply stellar idea here - Hixie's Natural Log: Content-Type is dead - browsers were broken, server configs are broken, so there can't possibly be any reason to use this header. The browser can't use it, so nothing else should either.

I think it may be time to retire the Content-Type header, putting to sleep the myth that it is in any way authoritative, and instead have well-defined content-sniffing rules for Web content.

You shouldn't throw away the Content-Type header even if server configs aren't easily controllable by the author. Go ahead and do the right thing - use the document context and tags as a hint on how to handle the content, use the content-type along with the content itself. There's nothing wrong with applications retrieving the resources referenced by an img tag to assume that the retrieved content is an image.
The only arguments people may have is when there is little context available when retrieving content (no hypertext source document) and the retrieved content could be interpreted in several ways.

Telomolecular Nanocircles

This description of telomere technology from Telomolecular looks interesting:

Synthetic DNA Nanocircles are a biomedical nanotechnology invented by Dr. Eric Kool and colleagues of Stanford University. These nanometer-sized circular DNAs have been shown to elongate chromosomal telomeres in vitro. They consist of DNA bases arranged in a sequence that templates the lengthening of telomeres by repeated addition of new TTAGGG sequences. Nanocircles have shown promise in telomere elongation in human tissues. By combining nanocircles with new proprietary gene therapy and delivery technologies, Telomolecular believes that nanocircles might work efficiently in living animals.

Google Base Storms Into Europe

From webpronews.com:

"The search advertising company will have a lot of catching up to do in Europe, where Amazon has relationships with businesses like UK-based Marks and Spencer. Google may have to pitch something beyond its online capabilities, though, according to the FT report:

One big UK retailer with no online presence said on Wednesday that Google's retail offer would be of interest if the internet company could also arrange for distribution. This potentially huge task has raised doubts about the long-term business models of other online retailers such as Amazon.com.

Doubts over Amazon's distribution don't have much grounding in reality. The company has built up its distribution network over the past decade and does have some knowledge in the area. Google's expertise at distribution probably does not go beyond making a change in a router's access control list and opening up a website. "

I like that last bit - Google's expertise at distribution probably does not go beyond making a change in a router's access control list and opening up a website.

The article goes on to suggest that Google team up with UPS for the distribution center. I wonder if that would work.

Google Base - Attributes

From a post by Mark Baker, I started looking at Google Base - they have a lot of interesting features - easy creation, bulk upload, extensible attributes and a
predefined set of attributes.
I've only looked at the tab-delimited bulk upload, but it looks pretty reasonable so far.

For listing products, the process seems pretty easy - http://base.google.com/base/a/1214338/9143072239359280355 . Their payment system is also integrated, so Google now has a real offer listing system with decent search. I imagine the next step is to syndicate the content as xml, wrap the data in a skin and provide good indexed search.

Service Oriented What?

Maybe I'm misreading this, but I almost fell over laughing at the post on Service Oriented Enterprise

3. REST, as the principle foundation of the Web, has proven that it too can scale - but the areas where it has proven to scale the best are related to the movement of unstructured HTML documents using a constrained set of verbs.

The but part cracked me up - from my viewpoint, it's the movement of unstructed documents and using a constrained set of verbs that allow for the scaling demonstrated by the Web.

Naked Answers

I like what Werner wrote about the visit from Shel Israel and Robert Scoble. Heh, wish I'd been there. I must have missed the email...

Anyway, this particular part struck me as being very true :

Amazon is a long time pioneer in the space of involving their customers with our product. And we really listen to our customers; any Amazon employee who encounters an issue on a forum or weblog or at any other place is empowered to escalate those issues internally immediately until they get fixed. Customer feedback is essential for Amazon and we will use all effective means to get it.

A few times over the past few years I've brought issues to internal teams and each time they have been totally positive about working on them - often it's a 'user education' thing (usually I'm the one being educated), but when it's a larger issue I've seen teams honestly re-examine what they are doing and incorporate the feedback and real change really happens.

Ripple - Internet IOUs

From a post by Mark Baker, the
Ripple payment system looks facinating! This is very intriguing because two nights ago I was wondering about payment systems and whether banks support 'request for transfer' requests in addition to just transfer-in and transfer-out requests. I imagine the problem of spam with request-for-transfer-out would be a problem. Anyway, the Ripple system is a distributed network of trust for the exchange of digital IOUs. These can be exchanged for products, services or even cash - anything of value. It'll be interesting to see what happens if someone integrates this with Edgeio or Craigslist (think 'mashup'). I haven't finished reading the whole site, so there are lots of open questions around fraud and actual overhead of participation, but this is all very cool.

Their About page is a good place to start.
From their FAQ:

How do I cancel debt on the system once it's been settled?

When two connected account partners settle a Ripple obligation outside the system, for example, by using cash, the person receiving the settlement must record it in the system by returning the settled IOUs to their issuer. They do this by making a regular Ripple payment to the person who settled their debt with you. Your mutual account balance will be changed to reflect the fact that the debt has been settled. In effect, they are using the IOUs they hold on the Ripple system to purchase cash.

The Internet is evolving it's own system of governing - of the people, by the people, for the people. First it started with communication, then community and now commerce. It seems inevitable that it's members will become an independent government unto itself. It won't start out perfect. Behavior, laws and regulations will continue to adapt, just as historical governing regulations have changed over time. But this tells me that there is a reason why this technology stuff is important. It is the technology of these Internet based systems that help private citizens remain private and remain citizens - and keep on being free. All changes start with a small ripple.

Irish pub on St Patrick's Day

On most Friday's, our team goes out for a lunch in Seattle's International district - mostly Asian, Thai and Indian food. But this Friday as we were heading out, we remembered it was St Patrick's Day and started looking for an Irish pub. The ones we knew about were kind of far, so we decided on going to F.X. McRory's - I always thought of it as a whiskey bar, but figured it had some Irish-ness to it.
When we got there, the staff all had some sort of green hat or shamrock painted on their cheek, so we knew we found the right place. There is a huge dining hall off to one side with high ceilings, old wooden floors and super tall windows letting in the sun - it was very nice. But the best part was that our table was right next to a band playing drums and bagpipes! I love the sound of bagpipes, but dang they were loud . We noticed later that the musicians had ear plugs - cheaters.
The menu had stew, shepherd's pie and corned beef and cabbage sandwiches - so I got the sandwich. I'd never had corned beef before and I have to say it's good enough try again someday - especially given a tall glass of Guiness to go along with it. But I'll have to warn you, the little dish of what looks like cabbage isn't cabbage, so if you take a big bite (like I did) you'd better really like horseradish.

ArtRage 2

If you or someone or know has even the slightest interest in painting, either on canvas or the computer, go right now and see ArtRage 2 - it's a fantastic program that feels like real painting. I really need to get a pressure sensitive tablet...

Pandora - find more music

Wow, the Pandora site that
plays similar music based on innate musicality is way cool! It's based on something called the Music Genome project that categorized songs on 40 or more dimensions. Fancy fancy!

"Ever since we started the Music Genome Project, our friends would ask:

Can you help me discover more music that I'll like?

Those questions often evolved into great conversations. Each friend told us their favorite artists and songs, explored the music we suggested, gave us feedback, and we in turn made new suggestions. Everybody started joking that we were now their personal DJs.

We created Pandora so that we can have that same kind of conversation with you."

Amazon S3 - finally, a real REST service

The new S3 Web service has what I would consider a very well designed REST interface that is also well documented (all the Amazon Web services have decent documentation). The current set of Amazon Web services is starting to grow nicely.

The S3 service is advertised as a Web-scale storage system with a super simple API and apparently very low storage and bandwidth costs. Each resource stored can be up to 5GB and also has support for ACL based permissions - to access the ACL, just append "?acl" to the resource. That's pretty simple and effective.

Microformats and product data

I've been hanging out on the microformats.org discussion list related to product data - too bad the email list isn't linkable. Anyway, these are my latest thoughts:

I am also wondering where and when a microformat for products would be useful. I'm generally wondering when /microformats/ are useful. (see http://microformats.org/about/) It seems to me that general narrative text and general conversations in text sometimes have 'hard' data like people, places, dates, times, phone numbers, etc Marking these as /being/ people, places, dates, etc. makes it easier for software to process - but the underlying data below the markup is still readable and sensible to people. Product data isn't like this at all. It's an unregulated mass of named properties with little rhyme or reason - there is no need to gently superimpose tags without breaking the elegance of the underlying data because there simply isn't any underlying elegance.

What it sounds like may be useful is a simple way for people to talk about things and reference those things, but without needed to fill in the blanks for unnecessary detail. For example, if I talk about the new mountain bike that I bought and immediatly scratched up by tipping over, it would be cool if I could tag the phrase "new mountain bike" as a reference to a product - but I don't want to actually describe the composition of the bike, the kind of tires and suspension and whatnot. People don't do that, manufacturers do.

If we want to solve this for manufacturers, then just use XML. Then the next problem is how can software determine which attributes from which manufacturers mean the same thing? That's where a shared 'open product attribute' database would be useful.

Gosling down on PHP and Ruby

This is a hoot - from
James Gosling at a Sun conference - narrowly correct but broadly missing the point :

'There have been a number of language coming up lately,' noted James Gosling today at Sun's World Wide Education & Research Conference in New York City when asked if Java was in any kind of danger from the newcomers. 'PHP and Ruby are perfectly fine systems,' he continued, 'but they are scripting languages and get their power through specialization: they just generate web pages. But none of them attempt any serious breadth in the application domain and they both have really serious scaling and performance problems.'

As a language, PHP and Ruby may not have the breadth of Java on a single machine - after all, they just 'generate' web pages. Well, they do a little more than that I suppose, they can handle incoming Web requests too. And what applications can you build and at what scale if all you can do is process messages over a network? I suppose the breadth of the applications possible is the same as the breadth of the network your messages travel over - which in this case is global. Maybe someone should tell him that 'the network is the computer'.