Computers

I jumped out of the Google Developer Day in Hamburg, Germany, because this is big news (coming straight from Sydney’s Developer Day): Google released an offline web app framework called Google Gears. One of the first applications is an offline version of Google Reader. Google’s Gears homepage explains that Gears is ...

<<... an open source browser extension that enables web applications to provide offline functionality using following JavaScript APIs:

- Store and serve application resources locally
- Store data locally in a fully-searchable relational database
- Run asynchronous Javascript to improve application responsiveness>>

Google in a press release states that Gears “marks an important step in the evolution of web applications because it addresses a major user concern: availability of data and applications when there’s no Internet connection available, or when a connection is slow or unreliable.” They go on to say that making the browser environment more powerful is increasingly important (Google snatched up a couple of Firefox developers – this makes even more sense in the light of this announcement).

You can install this for a couple of operating systems, though Google warns that this is an early-access developer’s release (which would be called Alpha, but Google calls it Beta). Like many recent announcements, this one too had been foreshadowed in Google’s internal documents which leaked last year. By moving web applications into the offline sector, Google Inc is moving into Microsoft Office competition territory even more clearly.

Gears is also free & open source, so Google is hoping to get the community to accept this technology and run with it. Without developers, Gears won’t be able to take off, which might explain the announcement around the time of the Developer Day. Already, according to Google’s press release, other big industry players are in the boat, including Adobe (senior vice president Kevin Lynch: “[t]he Gears API will also be available in Apollo, which enables web applications to run on the desktop”), Mozilla (chief technology office Brendan Eich: “a significant step forward for web applications”) and Opera (chief technology officer Håkon Wium Lie: “we’re excited to work with Google to extend the reach and power of Web applications”). Google says their “long-term hope is that Google Gears can help the industry as a whole move toward a single standard for offline capabilities that all developers can use.” Being so nice and open to turn a powerful online/ offline web framework into a commodity naturally makes business sense for today’s web application leader, Google; we saw what happens when they figure something doesn’t make business sense, even when it benefits developers.

TechCrunch’s Nick Gonzalez in the meantime reviews how Gears is already enabled for Google’s Feed reader:

<<Reader will add a green download button to the user interface. When you click the button, Reader will download the last 2,000 messages to your computer, preparing your computer to work offline or under a spotty internet connection.

Downloading will take place in the background, using the asynchronous JavaScript API. While offline you can read these articles and carry out your usual sharing and tagging. When you get back online, just click the button and Reader will sync your offline activity with their server. Right now the syncing is initiated manually, but it’s easy to see that it will become more seamless as the program develops. Gears could conceivably solve the large data overhead problems of Google’s AJAX applications, pushing updates to your desktop instead of slowing down your browser. >>

At the live presentation in Mountain View, Google suffered from the web OS equivalent to the Windows blue screen (you know, a browser error message) but this may have had more to do with the connection Google set up for the presentation than product bugs.

Ongoing comments

[Thanks Kevin G., Googlaxy and Krishnan S.]

[This post was updated as new information came in.]

[By Philipp Lenssen | Original post]

(InfoWorld) - The alpha geek consultants at the McKinsey Quarterly are out with an interesting new research paper that says app developers have to get lean – as in, adopt lean manufacturing techniques. That’s right, you know who you are – slim down, toughen up, drop and give me 20!

Of course, these are the same McKinsey folks who routinely helicopter into large corporations and charge millions of dollars to provide the CEO with a list of people to be laid off or “restructured,” as it were. So let’s remember that whenever they propose a strategic, 2-to-3-year organizational transformation of this type, they’d probably also like to get hired to manage it!

That said, the argument they make is provocative: that ADM (application development and maintenance) processes are poorly organized and fraught with waste and rework of the same kind that plagued manufacturing before the Japanese pioneered lean workflows.

Despite massive efforts to cut costs through offshoring, McKinsey notes, ADM (of which labor comprises 80 percent of costs) now accounts for half of IT budgets and continues to climb. Furthermore, there are big differences in app dev productivity among large organizations – for most, applying lean principles to app dev could increase productivity 20 to 40 percent while improving quality and speed.

Getting these benefits, according to the report, means thinking of ADM “as a kind of factory.” And then systematically eliminating sources of waste such as overproduction (unnecessary functionality), rework (requirements changes), “wasted motion” (ineffective prioritization), “wasted intellect” (limited developer cross-training), “wasted time” (idle developers), and inventory waste (maintenance backlogs).

The report runs through traditional “lean” solutions to each of these problems, such as “flow processing,” which “reduces overcapacity by aligning the rhythm of output with the flow of production.” Others include load balancing, greater standardization, segmentation of projects by complexity, and “quality ownership which extends beyond the testing group.”

And of course the consultants note that achieving these efficiencies requires not just process changes but shifts in behavior and new management tools (which they’d be happy to provide, no doubt). And they note that making these changes can be tough, because they have to become embedded in an organization’s culture to work, often “overcoming stubborn resistance” along the way.

Traditional software development methods such as CMM (Capability Maturity Model), function point metrics, and CASE tools fail where lean techniques would succeed, claims McKinsey, because they fail to address organizational and cultural issues such as alignment between business and IT – or the waste that can occur in the early definition and design phases of a project.

Does this mean I’d recommend dropping a couple mil on a McKinsey lean study for your company? Maybe, but first go out and spend $16 on a great book called "The Goal," by Eliyahu M. Goldratt. It’s a fast, fun read (there’s even a romantic subplot), and it gives you a real-world sense of what lean manufacturing principles are like in an actual factory setting (and will stimulate your thinking about how they could be applied to ADM at your company). Maybe buy copies for all your developers and managers. If that fails, you may have to call in the suits.

As you may know, tomorrow (May 31st 2007) Google will hold a developer day in different locations around the world. If you can’t make it, you can watch the developer day webcasts for Mountain View and London – and if you do make it, please add your photos and other impressions in the forum thread.

According to a Google press release, the webcasts will include:

<<- The Mountain View keynote by vice-president of engineering Jeff Huber (10:00 a.m. PDT)

- The London keynote by open source programs manager Chris DiBona and geospatial technologist Ed Parsons (4:00 a.m. PDT)

- Presentations by Google director of research Peter Norvig on artificial intelligence; Google technical director Mark Lucovsky on Google’s AJAX APIs; Google tech lead Adam Sah on distributing applications with Google universal gadgets; and Google Web Toolkit co-creators Bruce Johnson and Joel Webber on building AJAX user interfaces in Java.>>

And these are the webcast locations:

• Developer Day London
• Developer Day Mountain View Exhibit Hall
• Mountain View Room A5
• Mountain View Room A6
• Mountain View Room A7
• Mountain View Room A8

I wonder if any of the new technologies released yesterday – Street View and Mapplets – will also be covered in the different sessions, though there seems to be no hint in the session descriptions so far.

Ongoing comments

[Thanks Anthony H.!]

[By Philipp Lenssen | Original post]

Checking a server that installs a variant of Trj/Cimuz, I came across a link that pointed to remover.exe file:
 

After analyzing the code of the file, I noticed that it uninstalled the same variant of Trj/Cimuz that had been previously installed from that very same server.

I suppose this is the way the author uses to make tests in order to check if the Trojan works properly and then, get easily disinfected using the uninstaller.

(InfoWorld) - I should have gone to law school. If I had, I could right now be planning where to build my new summer home, 'cause you know I'd be on Microsoft's team getting ready to sue anyone who might be using a vaguely Redmond-resembling code module buried somewhere in one of my open source apps or OSes. Cha-ching.

Or I'd be on the Linux Foundation's legal team, probably with a beard and a corduroy jacket bearing elbow patches, now gearing up to countersue Microsoft for any Penguin-resembling code buried somewhere in a Windows app or OS. Ka-cha-ching.

Or I'd be on some corporate legal team getting ready to sue either Microsoft or the Linux Foundation for crippling an app my client considers critical. Ka-cha-chingeroonie.

But I'm not. Molecular LSAT scores are one problem. The other is having this inner need to work on projects that actually have the feeling of moving things forward rather than warring for a team that simply seeks to block progress in the interest of "competitive advantage" — no matter whose side I might be on. That's why I'm ignoring the impending Microsoft-Linux patent IP spat in favor of a late-day observation from last week's Interop: Office Communication Server (OCS) 2007 has telephony vendors worried. And I really don't get why.

Not because it's from Redmond — though that probably does add a spinal twinge or two — but rather, I think, because it's foreshadowing a technology trend. And that would be VoIP PBX software smarts moving off the box and onto the server. Sure, that means they might be losing some feature breathing room in the short term, but with your big-picture goggles on, it's easy to see it as an inevitability.

For example, unified messaging is pretty much a ubiquitous line item on VoIP buying lists now, but the depth varies. Voice mail as e-mail is basic. But the Office Comm product line is moving those expectations away from old phone paradigms and into new territory. Presence management, ink IMs, multiparty chatting and conferencing, the way-cool RoundTable conferencing hardware and internal as well as federated security. Just a few of the major line-item features in OCS — and I didn't even get to the new emoticons.

Handset vendors may be the ones doing the most sweating. After all, with OCS behind it, Office Communicator makes an amazing case for moving your whole business to the softphone model: a phone that can automatically pull up a CRM record in response to caller ID; go into multiparty videoconferencing at the touch of a button; let your boss know if you're really working and where; share apps and files right when you're talking about them. Yeah, handsets are going to have their work cut out for them.

But does OCS really spell impending doom for independent VoIP PBX and gateway vendors? Not all that much. In fact, it might give many of them a ready-made product road map. OCS is clearly one level above that kind of hardware. For one, there'll be a need to build OCS-capable appliance boxes. When the Small Business Server and Centro folks get off their heinies, a software/hardware bundle that can take you from an empty executive suite to a full server and communications hub in one install sequence might make a whole heckuva lot of sense. 

The same can actually be said for the handset folks, too. After all, wearing headphones and talking into microphones appeals to only one segment of the population. Cranky old diehards like me are going to miss their old fist-grippable handsets enough to make building smart, USB-connected telephones worth the bottom line effort. Sure, Redmond's sniffing around the edges of this market already, but that's hardly going to close the door on competitors.

Hey, the market's uncertain. But while that may generate fear when we're talking about outer space exploration, entrepreneurs should be embracing it. Either that, or think about law school.