CTO Blog

I think that Apple understands as one of the few that people want devices where you just push the button and it works, without needing to go through tons of manuals or getting the family's computer whizz kid to help you.

That's probably why game consoles are so popular, it "just works", no hassle with checking if your PC configuration is good enough for your game or doing messy stuff with device drivers.

Apple got it right. Look at their self configuring Bonjour network protocol, ease of use of their iphone/ipod touch, Time Capsule. Even my mother is a pro now with her Macbook with iPhoto.

Recently bought a new Mac (again) and an iPod Touch. I don't mind spending bit more money for these jewels :) Wished we got new Macbook Pros at Capgemini instead of our heavy and ugly Lenovo Thinkpads...

Cheers, Lee

Hi Lee,

Don't even get me started on this theme of Capgemini's own laptops, tears might come easily. Obviously, your statement about the esthetics of the Lenovo is completely biased and is most certainly not an official Capgemini point of view (oh, and you're right).

Guys, I think that's the main advantage of working for a startup : you get to choose your own laptop. True, peer pressure makes it always end up being either a MacBook or a MacBookPro - but nominally you're given the choice. And nobody would choose another computer in the first place anyway. Plus the black MacBook is simply too great not to be loved. Plus...

@Guillaume: I guess at Capgemini we could feel at least a little like a startup by using black MacBook Pro's.... No ...? And it happens to be exactly the new laptop I bought myself, for that matter. Don't you just love being predictable in this case?

You could try sneaking in some fuzzy logic in the mix along those lines :

"I'm a corporate blogger -> therefore I'm representing CapGemini to the community -> therefore I need to showcase our utter coolness in any occasion -> thus I'm entitled to a MacBook". Then once you've opened a breach you can start trying to exploit it ;-)

Re predictability - let's say that one might accept to admit that the concept may have its advantages in a set of given, particular and well-determined situations... (I wonder whether I'm not being a bit euphemistic here).

@Guillaume: interesting scenario! Pity that there is no way to show off as a blogger other than to put 'powered by a black Macbook pro' underneath every blog item. Would probably look just a little bit overdone.

The recent debate on whether IT should let users manage their own PCs is relevant here (http://ask.slashdot.org/askslashdot/08/04/02/1936207.shtml). Some organisations, such as Google and BP (http://www.infoworld.com/article/08/04/02/14FE-user-managed-pc_1.html), and now trialing giving users Technology budgets so they can buy and manage their own hardware. In my opinion, this is a lot more useful as it gives technology savy users more control of the software and hardware they use. As long as they can access enterprise applications there should be no issues with such an approach. It would also potentially cut down on helpdesk calls and modification requests, as the users would be responsible. Obviously, it would not suit everybody in the enterprise, but those savy enough should have the opportunity to do this. Perhaps it is something Capgemini could pilot?

@Richard: such a strategy will particularly - I guess only - work with more open standards and utility-based applications (to start with, all 'office'-like applications such as e-mail, word processing, spreadsheets and e-mail through a standars, ubiquitous browser such as Firefox). Capgemini benefited tremendously from the extreme standardisation (including the same laptop for everybody and one single image) of its thick-client/ PC infrastructure, and such a business case must always be maintained, even when we have high expectations of lap top users to manage their own computer, EVEN if that computer is a Mac...

It is interesting to see that almost any post about Macs spur of a lot of comments :) Being a devoted Mac user for a long time it is great to see that you have "seen the light" in terms of understanding what Apple wants to do with things like the Time Capsule. Usually comments around Apple products are so focused on analyzing technical specs and stating that things are missing and the product is overpriced and thus missing the whole concept of the product. In this case, to make people actually do backups by making it as easy as it obviuously needs to be. Apple did a study before this product and found out that only four percent of the users had automated backup solutions in place.

So what Apple usually sets out to do is think really hard on how to make things work in an easy and integrated way. Something that users not only can manage but preferably also like or even love. That is why style has been and always will be a part of the functionality.

Why do Apple succeed? To me it is a lot about corporate culture and Steve Jobs is of course a vital part of that. To many people today have adopted a "good enough"-ideology and Steve is known for his desire for perfection. I usually say that many "good enough" often results in a chain leading to rather bad.

There is an interesting article about the Apple way here:
http://apple20.blogs.fortune.cnn.com/2008/03/06/the-unfiltered-steve-jobs/?source=yahoo_quote

@Alexandra: don't worry, i've seen The Light already many years ago and am running an Apple-enabled household for quite some time now :) I even have a Apple Cube in my portfolio, which is now quickly becoming a collectors item. Your point about the backup is exactly what it is all about (although Steve's fine nose for downright coolness helps too): making the complex obvious and simple. I have lost a lot of content twice in my life due to a lousy backup routine, guess it won't happen again from now on.

Since Mac's Intel move, I've switched to Mac. The combination of Unix, a good UI, and the ability (Parallels) to run native windows applications if really needed, I just superb. Meanwhile we're even using a Mac mini to store/watch videos in the living room, and ... needless to say, some apple TV's are located strategically in the house.

Ron said: "@Richard: such a strategy will particularly - I guess only - work with more open standards and utility-based applications (to start with, all 'office'-like applications such as e-mail, word processing, spreadsheets and e-mail through a standars, ubiquitous browser such as Firefox). Capgemini benefited tremendously from the extreme standardisation (including the same laptop for everybody and one single image) of its thick-client/ PC infrastructure, and such a business case must always be maintained, even when we have high expectations of lap top users to manage their own computer, EVEN if that computer is a Mac..."

I do think this is a valid point. Presuming using Mac as the standard is not an option, one could still take a more differentiated approach.

1) For the Infostructure, do Indeed use open standards for corporate services. Aren't we telling clients to do so to? Some points of attention would be: exchange (calendar!) and follow-me printing [ the only corporate applications I have problems with when using my Mac ].

2) Adopt one desktop platform (for example, Lenovo + "the image") as the standard for "desktop help". In other words, if you're using a different setup, you're on your own for maintenance. The key condition is of course that Capgemini's infostructure uses open standards, and that documentation is available. This shouldn't have to lead to less efficiency. Furthermore, unless Capgemini would explicitly forbid people to use their own stuff, people will do so anyway. Better allow it, and manage it efficiently! Give people a choice: A desktop "allowance" + documentation, or a standard desktop.

Interestingly enough ... when I moved to Capgemini (at the start of this year), I started using the standard issue black box (which I have adorned with one of these apple stickers). However, since the start of this year I've had to re-install the image three times. After the third time I gave up, and inspired by another Apple die-hard within Capgemini, I switched back to my Apple again, with ever so more conviction.
I only use the black-box to sync Exchange's calendar with my Plaxo account, while patiently waiting for: http://www.snerdware.com/groupcal/ to become available for Leopard.

@Erik Proper: I wholeheartedly agree. For two years I've been using a privately-owned mac laptop (first powerbook, now macbook pro) within Capgemini. As a result, collegues wanting to use their macs find me for help.

Managing this initiative in a more structural way will in the end create happier consultants.

@Erik Proper: I wholeheartedly agree. For two years I've been using a privately-owned mac laptop (first powerbook, now macbook pro) within Capgemini. As a result, collegues wanting to use their macs find me for help.

Managing this initiative in a more structural way will in the end create happier consultants.

Back to the substance of the article (I'm coming to it at last ;-) )

One of the important things to remember is that this kind of invisible, almost perfect synchronisation and coordination comes at a cost : that of vendor lock-in.

Take the brand new Google Application Engine : at first look, a great, all-in-one tool that provides you with the ability to build and deploy applications in no time, without having to bother about making them go live, network & scalability issues or even user management. Heaven, right ? This means however that, you're depending on Google's DB and custom query language, on Google's ID management, on Google's bandwith, on Google's... While this might suit an individual developer perfectly well, it imposes a heck of a burden (read : potential risk) on companies that choose to use it.

The same goes for future iPhone applications. Sure, they're easy to build & easy to deploy thru a great SDK + the iTunes application delivery platform. That's cool for Apple, because it's the _only_ way you're allowed to develop and provide them to potential customers - meaning if it sucked nobody would be able to build legal iPhone apps. Oh and by the way, did I mention you'll have to give away 30% of your revenues to Apple for this ?

To sum things up : if you want to get your part of the pie in one of those dream dev environments where everything's running so smoothly, you're basically giving away a bunch of elementary freedoms. That's mostly fine when you're an individual storing your holiday pictures, 10 movies and a bunch of multimedia files, but does this really hold true for a third-party company trying to base its business model upon it?

Let's ask the market : most of the basic, mission-critical bricks of some of the biggest companies operating on the web (think Yahoo! or facebook here) run on Open-Source software. Apache. Tomcat. JBoss Middleware. Linux, Debian, MySql, Perl, PHP... What's their common denominator ? They're open and free for anyone to build upon and improve. Imagine you're relying on a database meant to handle millions of requests. You've identified this obvious memory leak. Your engineers even came up with a way to fix it. Problem is, the database is proprietary and you'll have to wait till the end of the next release cycle, in 2 months. Or you can use MySql, fix the issue and have the whole community benefit from your work days later.

Sure, you've lost that nifty, high-level profiling feature the proprietary software offered. But your system runs faster, more stable for longer periods of time. Make your choice :-)

Companies are looking at the tradeoff between freedom & integration from a different point of view than their consumer's. The "Invisible Infostructure" sure sounds cool, but the way it will be built and delivered to companies is crucial to its success. Companies such as Amazon might have striked the right balance already : specialize yourself in the infrastructure - but the infrastructure only, make sure it runs fine but do not try to compete with your potential users. I'd rather run against fellow startups runnig on Amazon's EC2 than against Google playing on its own turf.

The remaining question is: why couldn't companies benefit from openess + freedom and integration + synchronisation at the same time ? There's no fundamental reason why things should be so, and there's a technical answer available right away : interoperability. Industry-wide agreements over Open Stadards offers room for freedom and integration to work together closely. Don't give up on either one : you've got no reason to do so when you can get both !

@Guillaume: well, this is certianly opening up the discussion in a slightly different direction. Personally, indeed I think 'inter-operability' is as much under-rated as 'open source' is over-rated. I just don't like the idea of being able to tweak into my own database system if it doesn't perform well enough or contains a bug. Unless for the fun of it...
If there truly is an Invisible Infostructure, let the supplier deal with it. I could certainly live with the idea of being locked in into Apple, as long as the infrastructure is smooth, managed and feels natural. And - of course - it needs to provide me with the interoperability features I need to base my TRUE business on, which most certainly is not running and optimising my own infrastructure (unless I am indeed Amazon, Google, Capgemini Outsourcing, IBM, etc.).

Thanks for the excellent builds!

Couple of comments from an outside reader - on the side issue of laptops, I've used both Mac and PC laptops, and don't mind switching back and forth precisely because of the mostly-invisible infrastructure of the web. For a desktop, though, I still prefer a PC environment just for price points & flexibility. But the original issue, I see Guillaume's argument in favor of open-source, but it's all predicated on one overlooked condition in his post: "You've identified this obvious memory leak. Your engineers even came up with a way to fix it." But that precondition is exactly why a multi-billion-dollar software industry grew up in the first place - why do I want to have to identify a memory leak or have *my* engineers internally come up with a way to fix it? That's what I want S/W companies to do, with expertise, rapidity, and reliability. Yes, you can often get the first two of those 3 characteristics from "the crowd" of open-source developers, but not reliably. If I have mission-critical systems, I want a reliable patch pronto. I don't want to have to staff out a team of multi-talented engineers in expectation of identifying & fixing every problem imaginable; I want to know that Oracle or SAP or Microsoft or Apple will likely identify it first and fix it - and if necessary remote-patch it for me. Just my 2 cents... my goodness, this post certainly sparked a variety of thoughtlines! Thanks.

@Lewis: Guess you're definitely right on the 'I can fix it myself' thing around Open Source. That is just not really relevant to most of us - or it shouldn't! Long live the Truly Invisible Infostructure. That said, interoperability (and through it, invisibility) in many cases is driven by activities in the Open Source community, and we should carefully watch. Would be interesting to hear from you - after all, you are a Microsoft CTO - how you see this.
Difficult indeed to get back to the original 'Apple' subject, but then again you linked to a - well - interesting idea of a colleague for a new Microsoft acquisition target:
http://lewisshepherd.wordpress.com/2008/04/17/what-if-microsoft-bought-a-slice-of-apple/

Thank you for even mildly suggesting this. Might ruin my weekend :)