General “Must Have” Requirements for the next “Killer” Web Service

Just my opinion, your thoughts welcome

1. Must have Twitter integration, in and out
2. Must be visible across all web services, using RSS as the universal glue
3. Must accept OpenID, if not also be an OpenID provider to boot
4. Must be SEO friendly, if not a downright SEO “juicer”

In short, openness, visibility, and portability are necessary and sufficient for its success.

My idea would be for an “open” data store based upon OpenID trusts to extend personal profiles across web services. Yeah - I know there are already groups out there trying to grok this in group thought.

Someone just needs to do it and let the market decide. The internet has enough committees.

Hate to think what Twitter would look like if designed by committee… or community for that matter.

Branding and Esteem

Just posted an article on my Facebook profile about Dell’s poor quarterly earnings and how they blame their poor performance on everything EXCEPT the root cause of their troubles - terrible customer service. While Dell has reasonably priced and well equipped laptops and desktops, they are losing mind share and market share to Apple (yes, Apple).

Apple’s product offerings across the board are overpriced and underpowered. And yet - they are expanding in all areas: music downloads, music players, desktops, phones, laptops. They are on a tear, no doubt.

Why is Apple prospering, and Dell (and Sprint, another notable customer service story from hell) floundering?

Because Apple respects their customers. Apple products are an esteem issue with their customer. Customers are willing to pay a premium for ease of use, ease of hassle, beauty, elegance.

The marketplace of commerce and ideas is fast changing, and old thoughts of ignoring customers because you CAN are long obsolete.

Just ask Dell and Sprint. While you can.

Do As I Say, NOT as I Do

Ajaxian had a post today about how Firefox 3 uses an undocumented API call to get the same performance characteristics as Safari on Mac.

This kinda “private vs. public” debate is as old as the PC. Peter Norton became a millionaire by exposing the hidden / underdocumented bits of the bios, fat, and msdos, and Andrew Schulman went on to become wealthier - if not persona non grata at Microsoft - for writing about Undocumented Windows API calls.

I had an app that was in production for 18 years (YEARS) that used frowned upon direct video memory writes for fast display (which, in the day, everyone did writing commercial DOS software).

To sum up: its OK for Apple to use their own internal API calls to get a performance gain over competitors, but it is somehow “bad” if Firefox does the same to level the playing field.

Yeah - THAT sounds fair.

Using undocumented API calls is sometime the only way for commercial software written by third-parties to compete in a marketplace where OS publishers use undocumented APIs as a way to outperform those locked out of the Sanctum Sanctorum. Microsoft has been notoriously beaten up in the past over such shenanigans (”DOS isn’t done until Lotus won’t run” went an infamous apocryphal saying back in the eighties).

Apple probably isn’t getting as much heat now as MS did way back when, for a few reasons. Apple fan boys and gals usually question nothing that Steve and Co dish out. Most developers today are far removed the vagaries of OS API calls (or the OS for that matter) and really don’t understand the implications that using undocumented calls to gain competitive advantage give an OS publisher writing apps as well (Microsoft famously touted the imaginary “Chinese Wall” between its OS team and the applications teams).

The only real danger in using undocumented API calls is that when the OS changes and the API behavior breaks in some way, the OS publisher is morally under no obligation to notify of the breaking changes. It’s caveat emptor, baby.

Not even as bad as the risk people take when they hack the SIM to use carriers other than AT&T on their iPhone.

In short, if using undocumented API calls levels the playing field, do it - with your eyes open - because if it were REALLY that bad, the code wouldn’t be in the WebKit to begin with.

Reliable Services

One bane of the social web is the unreliability of services that we depend upon more and more to extend our brand, connect with friends (real or imaginary), or flat out do our work (work? what’s that?).

Twitter (at least for me) is more noticeable when it is down, as it is one of the more immediate and impactful of all the social web services. I update all of my sites that have statuses using Twitter integration in one form or another. I regularly use 3 twitter clients a day (two browser based, one mobile based) to check tweets and update my status.

When Twitter is working, it is the best thing since sliced bread. When it is not… ah, there’s the rub. It is down ALL the FREAKIN’ time.

On the one hand, how can one complain about the uptime of a “free” service? And yet, how can I delegate more trust and business function to a service that is as unconstant as the moon (nod to Will Shakespeare)?

I’m picking on Twitter in this regard, though there are plenty of other examples out there in Web 2.0 land (the Bloglines Plumber and the LinkedIn Wizard - who made an appearance last night, BTW - are just as recognizable as the Twitter Cat and Robot screen). Bebo had some ungawdly downtime number, measured in the double digit HOURS for a two month period, and that just ain’t gonna cut it ongoing.

The relative unreliability of social networking services, free though they are now, are real stumbling blocks to wider adoption by businesses that need reliable up-time “always there” access. Web services should STRIVE for the level of service and reliability that we expect from our phone service, water systems, and electrical grid.

Strive is the key word.

Think I am overhyping this? What if you were trying to alert a campus emergency ala Va Tech with Twitter or a service like it, and got the Cat and Robot screen? What happens when Skype goes down for a day?

Real world high availability systems are just as vulnerable as the virtual systems I’m ragging on. Just ask the people on the East Coast of Florida about their half-day blackout a couple of days ago.

Again, we must strive for a higher level of service from the online services to which we are devoting large chunks of our lives and intellectual capital.

Now, forgive me. I have to go interact with real people until Twitter allows me to chat with my imaginary friends.