Yahoo Mail Supports Free IMAP … But Only via Mobile

It’s hard to know what to say about Yahoo at this point. They have a lot of page views, but their internal organs seem to have been liquified from the inside by some flesh-eating plague.

Barely alive, they are of course responsible for their own continuing undoing. A great example is the failure to do something useful with the email user base. The purple crew has missed every opportunity to exploit the email product, from Google-style ads to Facebook-style social networking (back when they had a chance). And there’s the infamous CEO with her Napoleonic airs.

But I digress.

See, here’s the thing. On the Internet, unless you’re lucky enough to be a telco, you actually have to compete. Which means when Google gives away IMAP-enabled mail, and Windows Live gives away POP access and 25+ GB of online storage, you need to rethink your strategy of making your mail product harder to use while praying that the nine people on earth who haven’t heard of Google will sign up for your deluxe mail service (now with undercoating!)

But Yahoo has never been able to think one thing at a time. It’s not so much that they excel at multitasking than that they are schizophrenic. So while one group is committing ritual suicide with the email product, another realizes that to reach the mobile market, exposing mail via open protocols might help.

And hence they offer endpoints called something like imap.mail.yahoo.com and smtp.mobile.yahoo.com. Before you get all excited and ask me to check the hostnames and port numbers, though, I’ll drop the punchline: it appears Yahoo filters access to these services based on IP, and opens the service to mobile carrier address blocks (and carrier proxy addresses).

After seeing my phone successfully configure itself for these services, I couldn’t resist trying them from a desktop client. No joy. I guess if I were really hardy I could grab the public-net-facing IP address for my phone (by having it connect to my own server) and pretend to be that IP. But … really … is it worth it? No… with Yahoo still living in 2002, I’m afraid it’s just not.

Work 2 Different Logons (or Sessions) with Private Browsing Modes

Here’s a trick that’s fairly obvious to web developers but probably not so for everyone else.

Private browsing modes – such as Incognito (in Google Chrome) or InPrivate (IE 8) clear out locally stored browser cookies (snippets of tracking data) each time they are run, and do not share cookies with the main (non-private) browser tabs.

Since these cookies are the core mechanism by which web sites associate independent browser tabs/windows with a single user session, the private browsing modes can be used to interact with multiple different logons or accounts at the same time.

How does this work in practice?

Let’s say you want to interact with two totally different accounts (logons) at, say, Gmail, Yahoo, eBay, Orbitz, or some other site. You’ve probably noticed that if you are logged in to, e.g., Yahoo with one browser tab, and you open another window or tab in the same browser, and go to Yahoo, it will “know” who you are and allow you to interact with the same account.

In many cases, this is a desired behavior. But if you want to work with two different Yahoo accounts at the same time, it is quite tricky.

Here’s where the private browsing mode comes in.

Open a private browsing window/tab, and log on to Yahoo with the second account. Since the private tab doesn’t share cookies with the main tab, you now have two interactive sessions with the two different accounts, and they “stay separate.”

There are three catches though:

First, since the private browsing tabs are designed to discard cookies on shutdown, they will “forget” your logon when you close the browser even if you select “stay logged in” on a specific web site.

Second, it is possible (though unlikely) that this mechanism could fail if the site uses “Flash cookies” apart from regular cookies.

Third, while this trick works perfectly well, do not assume that the service you are connected to has no idea what you are doing. They probably don’t care. But if they do care, assume that they could guess that the same user was on both sessions (through IP addresses, NAT port assignment patterns, etc.) at the same time.

Two More Small Tales from the IP Front

There are a small number of areas in which I am an expert, and intellectual property is not one of them. For expertise, I like to refer folks to Mike Masnick and Techdirt.

Nonetheless, it’s hard to move through the technology world and not get bonked on the head by IP absurdities and incongruities calling out for some solution aside from lawsuits and Nancy-Reagan-style denials.

First up is the tale of two Comcasts, and two identical HD streams of Battlestar Galactica. One of these streams comes in via the HD DVR, and the other comes in via Bittorrent.

Actually, I’m leaving a bit of information out in the setup. See, the two HD streams from SciFi channel should be more or less identical modulo commercials. But they’re not.

Due to bandwidth allocation and network management issues, the “legitimate,” paid-for SciFi channel stream is full of bitrate-spike artifacts reminiscent of late ‘90s web video. So the HD channel, the HD DVR, the HDTV gets you … a pretty awful picture the second characters start fighting, running, or blowing things up.

Then you have the Bittorrent stream, ripped from SciFi, and carried in via the same Comcast coax line. This stream looks great, and makes a customer glad to have Comcast and a HDTV. In some strange respect, this latter stream may represent a problem in the “industry’s” opinion, even though it’s clearly what the customer wants and is paying to view, and even though the bandwidth (both instant and aggregate) is a small fraction of that required for the “broadcast” stream.

Next strangeness is Grooveshark, a free, crowdsourced, on-demand streaming service that appears to live on the razor-thin edge of legality if it has a claim on legality at all.

That said, the ability to send (or post/tweet/blog/…) a link that goes directly to a particular song is an extremely potent way to virally spread the music you like. And more music loved by more fans is the core asset base for any music industry, whether it resembles the 20th-century “legacy” record industry or not.

Without Grooveshark, there is always YouTube for sharing instant, no-membership, no-login-required tracks. But unlike YouTube, Grooveshark is structured in a way that encourages more exploration of an artist, album, etc.

So anyone wanna throw into a pool for how long ‘til this service gets shuttered? Wait, that’s not quite legal either. We’ll “just say no” and it’ll all just go away.

Suck Us All Into the Machine: Build FluidDB on Twitter and #Hashtags

I was reading today about a fairly amorphous, tag-based, public "database" concept called FluidDB.

To avoid the usual distractions, I will assume my audience knows all about relational and non-relational datastores, implications of tag-oriented metadata, etc. In fact, since my friends know about this stuff and they are most of my audience, I think I'm not too far afield making this assumption.

One reason I found the FluidDB concept interesting is vanity: about four weeks ago, I spent some time considering building just this kind of database ... on top of Twitter.

Why would one want to do that?

Precisely because Twitter already has a lot of meaningful data curated by humans and tagged with a well-known metadata scheme (hashtags).

In exchange for having very small or hypernormalized data records (since each atomic entry is limited to 140 characters minus the tags and any indexed keys), we get a strange merging of human- and machine-readable data.

Humans could read (and follow, search, etc.) data entities of interest.

And clearly the "goal" of automated (machine) participants (clients) would be to understand as much of the human content as possible, treating it as objects, tuples, logical inferences, or knowledge base "facts."

Moreover, the originator of a tweet, as well as any @-referenced recipients, are critical metadata. They are, actually, tags themselves in way which is linearly independent of the hashtags. That is, a from-@adbreind tweet (entity) marked #database is different from a to/ref-@adbreind tweet marked #database, while #database must be considered to (possibly) have a different sense than it does in, say, a from-@headius tweet. However, the same tools and semantic analyzers can be applied, essentially treating the writer and target of a tweet as special tags.

In this way, our twitter discourse, short enough to make machine understanding tempting even when the packed cultural references make such understanding impractical, merges us into the database and makes us "just another part of the machine."

AT&T “Fixes” My Phone by Downgrading It

I don’t want to be another corpse in the AT&T pile-on. AT&T has many problems, and they impact me, but let’s look at another technical clue about the network trouble:

I stopped in Tuesday at an AT&T store to see if they had any advice or fixes for the network issues plaguing my phone. The only thing the service rep said he knew of was to replace the SIM card. Sure, why not? Can’t hurt anything, can it?

Well the answer to that question apparently depends on whether ditching 3G access counts as “hurting.”

Since I got the new SIM, the phone spends most of its time on the EDGE and GPRS service, and only rarely gets a fix on the 3G service even when I’m in known-good 3G areas.

For a fleeting moment, I thought about taking the phone back and asking AT&T to do something to restore my 3G goodness.

But I quickly realized that the network’s inability to keep the phone on 3G was a big part of the problem. In other words, as so many iPhone users have discovered, staying on EDGE is better than flitting between EDGE and 3G and getting nowhere. Likewise in the case of EDGE and GPRS.

So for now, I’ll take my retro GPRS and its 50kbps or so of throughput. Since the setup and teardown of the data connection, as well as the latency in repeating a request, are the slowest part of the process, I’d rather use dependable 50kbps connection over an unstable, unusable 500kbps one.

Until I get a network that actually works, that is.

Link: An Illustrated Brief History of Augmented Reality

Check it out.

Plus who knew that Philippe Kahn invented the camera phone?

Apparently, a lot of people, just not me. The guy is so awesome.

Microsoft’s Real Punishment is Having to Fight with the Gloves On

By the time the Microsoft anti-trust cases wound down, it could be argued that the alleged damage was long since done to the industry.

In a curious reversal, it appears that Microsoft’s direct penalties from the case were only a tiny beginning.

The real penalties Microsoft pays are not in money, nor in shipping “K” and “N” SKUs. They are in product strategies unpursued because they would be too provocative. In other hands, such strategies  might be reasonable if aggressive, but for Microsoft they might look like “relevant conduct.”

Today’s announcements about Office 2010 (“14”) were … well … let’s just say if you didn’t read about it, the most interesting thing you missed was seeing critics point out that Microsoft, inventor of AJAX (literally, for the Outlook web client), is only bringing other apps online 10+ years later.

The reasoning here is not about selling client OS licenses. Microsoft could have moved to the cloud richer and faster and more profitable if it could take the gloves off.

Need to get Silverlight penetration up from its abysmally low numbers to where it can really compete with Flash and become a meaningful platform? Just ship it with Windows and make it a priority update to every Windows box in the world. Problem solved. Now we can get down to the real work writing apps. Or at least Adobe could actually face some competition. But Microsoft doesn’t dare do this.

Why not stitch cloud storage directly into the OS? I hate leaving files “on the other machine.” Right out of the box, anytime I see an “Open…” or a “Save As…” dialog box on Windows 7, I would like the default destination to be a secure folder on Windows Live SkyDrive. Using the provider pattern, other vendors could offer a similar service, and the end user could choose. But Microsoft doesn’t dare with this either.

It is quite possible that the long-term benefit to the industry of having Microsoft thus restrained far outweighs the “lost” value we could have had from Redmond. But let no one fool himself into believing that what we see from Microsoft these days is everything they have to offer.

VncSharp Rocks for Programmatic and Interactive Remote Access from .NET

VncSharp is a C# implementation of the VNC protocol together with a handy visual Remote Desktop .Net widget.

At first, the open-source VNC remote-access solution might seem like a surprising item to need in a Microsoft-based solution.

But once he gets you in the little room, VNC starts telling you that, in exchange for lower performance (than Microsoft’s own RDP), he can get you more flexibility, more features, no licensing issues, and access to remote Macs or *nix hosts. After these arguments (or is it the heat?) VNC looks a lot more persuasive.

Throw in the fact that you can run your VncSharp-enabled apps on Mono, and … well, it would be cool if that bought you a lot. But actually if you aren’t already focused on a Linux solution then the Mono angle is just another bullet.

VncSharp itself, though, works extremely well straight out of the gate.

When you see the documentation page – where the author essentially invites you to read the source to figure out how to drive it – you may be concerned. Or even start to form silent curse words with your lips.

Do not let that stop you. There are demo/sample apps that will show you what you need for basic use cases (e.g., popping open a “remote help” window will only require a line or two). And the source, if you need it, is elegant and straightforward to navigate.

A programmable VNC client is, perhaps, a niche product. VNC on Windows maybe more so. So it’s gratifying to see such a mature and streamlined OSS effort.