Monday, December 31, 2001

Good Easy Startpage A good, easy start page makes using the web faster and easier. Here's mine. Since I use google all the time, I've built a google query form right into the page and I've set Netscape such that opening a new window automatically goes to my good easy startpage. So searching with google is as easy as "special-N", search term, "enter". After the search form I have frequently used links.

You can build in any search form, and put in your own favorite links. Save the document as an HTML file on your local hard drive and set it as your browser's home page. Happy Holidays!

Sunday, December 30, 2001

2002 for Open Source Tim O'Reilly has a great piece on what's new for GNU/Linux and Open Source. His key topics: interoperability, open data, and freedom to innovate.

Tuesday, December 25, 2001

Happy Holidays! I'm taking a break for the holidays, so winterspeak will be pretty quiet until the New Year. But I do have a Christmas present in mind for readers that I hope to put up soon. Here's to a prosperous 2002!

Friday, December 21, 2001

Napster, and the Brand Online It's funny hearing execs at MusicNet and PressPlay talk about online "brands" (including Napster's). It's clear they don't know what they're talking about.

Online, customer experience is the brand. The Napster customer experience involved:
1) Going online
2) Finding any song you wanted
3) Easily downloading that to your computer
4) Listening to it as many or few times as you wanted
5) Putting it on different media

This experience is now owned by Morpheus and other file-sharing networks. The horrible, crippled experience MusicNet and PressPlay offer don't come close to this, and the post-Napster Napster won't either. The name "Napster" holds no value to the customer, only the experience matters.

Everyone knows MusicNet and PressPlay will fail. Better alternatives exist, namely 1)Morpheus and 2)CDs. I'm pretty sure the Recording Industry cartel is wise to this too, but they're not worried. They reckon the courts and Congress will help stamp out online file-sharing and are quite comfortable with the status quo CD sales represent. So MusicNet and PressPlay don't need to succeed, they just need to make the Recording Industry look like something other than a cartel colluding to restrict quantity until they digital rights management is built into all hardware, courtesy of a Congressional edict.

Thursday, December 20, 2001

Desktop is safe I just utterly failed to install Redhat 6.2 on a whitebox I put together for my dad (PIII, Win98, 20G hard disc). I know you're all wondering why I was installing GNU/Linux for a 50+ year old who mostly just surfs the web and emails--it's because he just surfs the web and emails, two things GNU/Linux can do fine. I expected stability and speed to be better under GNU/Linux, and the system could run (without upgrading) forever.

The speed requirement meant I didn't want to use a partitionless install (coupled with the fact that it makes GNU/Linux default and requires a disc for Windows--my dad would lose the disc and it would be all over). So, I'd need a LILO and a partitioned hard disc. The Official Red Hat Linux Installation Guide was utterly useless in this regard. It mentioned automatic partitioning, which sounded great except I couldn't find it. Manual partitioning suggests making /boot, root, and swap, without going into any details, and leaving out all the other partitions the installer itself recommends. At this point, not wanting to screw up a barely stable system, I quit.

I wish a team of engineers from Red Hat had watched me try to install their distro. Actual observation of user behavior would have made all the errors in the Installation Guide utterly obvious. Direct, non-directive observation of actual users struggling with actual technology products will do more to make tech useful than a million technical upgrades or marketing feature requests.

On the plus side, my experience with the actual local GNU/Linux community was extremely positive, as always.

Wednesday, December 19, 2001

The Acid Test for Intellectual Property Once upon a time, people understood that society benefited most when ideas (know in content production circles as "intellectual property") are in the public domain. Since they cost nothing to share, their market is efficient when they're free for everyone to use. In economic terms, ideas achieve Pareto Optimality in the public domain.

It was also understood that some ideas take up front investment to produce. Drug development costs millions of dollars. Novels take hundreds of hours to write. For idea producers to fill an intellectual commons, they must recoup the cost of their investment (otherwise they would never make the investment in the first place). Copyright law empowers the copyright holder to artificially restrict the quantity of the idea (keep it out of the public domain) and so make excess profit on it, this excess profit covering the initial speculative investment.

Therefore the acid test for good intellectual property law is obvious: does this law get ideas into the public domain as quickly as possible?

Let's look at how some copyright laws stack up:

1) Copyright for 14 years. Initially, copyright was only for 14 years (and limited to maps, charts and books). Since producing a creative work takes significant investment, it makes sense to grant the producer some monopoly power over that idea so they can have a return on that investment (and so invest more). 17 years is probably too long in today's world, so something shorter, like 3 years might be better.

2) Copyright for author's lifetime plus 70 years (retroactively extended). This clearly does nothing to enrich the public domain. Extending copyright beyond author's lifetime cannot possibly motivate him to produce more, because he's dead. Neither can retroactive extension, as the idea is already produced. 70 years is criminally long.

3) Patent on drugs. Drugs cost millions of dollars over several years to develop. If they were not patented, this investment would clearly never happen and we would have fewer drugs. This is bad, so drug patents are a good idea. How long should the patent be? For expensive things like drugs, probably a while, and for cheaper things, less time. Whatever it takes to support reasonable investment.

4) Patent on software/business models. Producing valuable software processes or business models must help a company make more money (otherwise they wouldn't be valuable). Since companies try to make more money anyway, these patents just keep ideas out of the public domain. They do not help create new ideas, and so fail the acid test.

5) Voluntary GPL: GPL'd work is copyrighted (well, copylefted) and not in the public domain. Nevertheless, it has many of the benefits of public domain work and therefore is clearly more valuable than proprietary software. Under reasonable copyright, GPL'd work would enter the true public domain in about 3 years. Under current copyright law, since RMS is around 40, Emacs won't enter the public domain until 2112.

6) Mandatory GPL: Software incurs development cost, and the GPL does not allow the producer to make up that cost. Therefore, if all software had to be written under the GPL, we would have less software, which would be bad. This is not a criticism against the "viral" nature of the GPL, which is the only tool defenders of the intellectual commons have to enlarge the public domain in today's oppressive copyright climate, but a criticism against the position that all software must be GPL'd. GPL'd software is better than proprietary software, but proprietary software is better than no software at all (Free Software folks disagree with this last point).

It's not difficult to think intelligently about what copyright law should be (even though the patent office, legislature, and legal profession continue to misunderstand its purpose). Copyright should be a friend to the public domain, and help enrich it with ideas it would not otherwise have. The acid test--"does this law get ideas into the public domain as quickly as possible?"--refocuses the debate on what an idea is and what role it plays in an innovative, efficient society.
Link to this article

Monday, December 17, 2001

Better Licensing Experience As Microsoft cracks down on piracy and builds draconian rights management into all its systems, it makes the experience of using that system worse. The GNU/Linux community is right to applaud such moves, because it increases the experiential edge Free Software has over proprietary software.

Sony understands modchips Sony is clamping down on modchips--dongles that enable playstation 2's to overcome regional encoding. Sony claims modchips (which aren't illegal) enable people to use pirated games--a serious problem because that's where they make most of their money. But even if people weren't buying pirated games, the chips overcome regional encoding, which reduces Sony's market power and therefore ability to price discriminate, which damages margins. Since overcoming regional encoding is quite legal (except in the US under the DMCA), Sony can't admit that's their real motivation--thus the useful "pirate hacker" ruse. Personally, I'm just glad to see them acting in their own self-interest for once.

Saturday, December 15, 2001

Updated links page Check out updated links.

Friday, December 14, 2001

Nokia vs. Microsoft Microsoft dislikes divided technological leadership, and since it now views all digital information as fair game, this means it wants Windows in all mobile phones. Nokia and company saw this ages ago, which is why they formed Symbian, a consortium dedicated to producing a non-MS OS for phones.

The press and investment community has been blathering on about cellular handsets getting commoditized away like PCs, but I think this is unlikely. Essentially, handsets are squeezed between two forces: carriers and operating system. In Europe, the GSM standard and swappable SIM cards in Europe mean that carriers have essentially been commoditized away--at least to the extent governments will let them be. Users are free to switch from one carrier to another, and to be honest, so long as coverage is reasonable no one cares who the carrier is. Complete interop (both voice and SMS) also means there are no demand side externalities, so the handset market should do fine with multiple major players. On the OS side, both embedded GNU/Linux and EPOC (Symbian's OS) represent good commodity OSes (without platform strategies underneath them) that are preferable to getting into bed with the Beast. Finally, although Microsoft might want to believe otherwise, there's little tying cellphones to PCs or servers--voice and SMS run on non-IP networks.

In Japan, the government has the entire vertical sown up courtesy of DoCoMo, and the fragmented standards in the US means that carriers have commoditized away handsets and OSes. The .NET based services Microsoft hopes will loop in phone users don't mesh with customer needs, and will fail. No market demand AND no way to leverage established monopoly means no place to play for Microsoft.

Of course, this doesn't mean that market saturation might not drive down Nokia's envious 15% margins. As Ollila looks for more revenue, it'd do better selling hardware with gold-plated integrated software on commodity OSes.

Wednesday, December 12, 2001

Dubai GNU/Linux cluelessness I spoke to someone from Dubai's E-government initiative and asked them what their systems would run on. They said "Microsoft is the Internet so we run on that." *Sigh*. Folks I've spoken to who use this e-government system moan that mandatory subscription fees make it more expensive than doing things the old way. At least we now know how to make money online.

Dubai was like Hong Kong piracy-wise until after the Gulf War, where Microsoft put pressure on the local government to stamp it out. Looks like they did a fine job, when I called a white-box assembler yesterday, they refused to pre-install GNU/Linux because they "didn't offer any pirated software." *Double sigh*.

All your bits are belong to us Microsoft does not see itself as a monopolist because the desktop computer is just one small part of where data resides, and Microsoft wants to control all data. Microsoft has patented digital rights management at the operating system layer, which means no one can produce digital rights management OSes if Redmond doesn't want them too. Moreover, under the DoJ "settlement", Microsoft is protected from ever having to disclose information about it's operating system pertaining to digital rights management. Ooops--that would be the whole operating system now. Finally, under the Microsoft backed SSSCA, NOT having digital rights management built into the hardware and OS layer is illegal. So you can't run GNU/Linux, or anything that isn't Microsoft, even if you want. Please write to the DoJ. (To: microsoft.atr@usdoj.gov, Subject: Microsoft Settlement)

Good, Easy Email Email may not be the most glamorous part of the Internet, but it's probably the most important. Back when Jeff and Donna was putting together the Palm Pilot, the average Silicon Valley exec was getting maybe 3-4 emails a day. I'm not a exec, but I now get about 200-500 emails a day, which is pretty common for tech folks. My average business school colleague gets 30-40 a day and feels overwhelmed by that.

When normal people start getting email overload, they just unplug, and limit their use of the medium. By contrast, tech folks come up with complicated schemes involving HTML, mediating interfaces, web based personal message spaces etc. etc. etc.

In my experience, more technology is rarely a solution to too much technology. Adding extra layers that complicate email will just make it less flexible, less powerful, and less easy to use. This is why HTML email is bad, and Outlook, with its poor support of plaintext, makes email less useful.

A while ago I wrote how Mark Hurst's Good Easy, a system of five simple applications working together, combines the power of Unix and the ease-of-use of a GUI to create a simple and powerful way to work over a network. In this context, the email client just streams bits in and out of the computer. It is not a repository for information, and therefore should be kept empty.

Here's how to empty your inbox (and keep it empty)

1. Put the oldest, crustiest emails at the top. This will help you remember to get rid of them.
2. Delete all spam (without opening it).
3. Go through all personal messages from friends and family (the most important emails). Read them, enjoy them, delete them. If you must keep them, copy and paste the text into a text editor and save it in your hard disk as "'initials of sender' 'date' 'keywords'".

4. Your inbox now contains email of middling importance only. Sequentially engage each email to remove it from the inbox:

- If the email is a to-do item or appointment, copy and paste it as a to-do or appointment into your calender program. To-do items should be tied to specific days, so Outlook users are out of luck. Delete email in inbox

- If email is a item of correspondence tied to a project, save it in the folder for that project as " ". Delete email in inbox.

- If the email is reoccurring (newsletter, list), read the email and enjoy it if you have time, else delete it. You'll be getting a new one tomorrow anyway.

5. Continue until your inbox is empty, and keep it that way.

Why this works

- Storing bits in an email client splits your file system, making it harder to find anything. By storing everything on the hard disc, a search (using "finder" on Macs, and whatever on PCs) can look through all your data at once.

- A cluttered inbox makes email useless. It's impossible to find, act on, or reference workflow if it's buried under 200 other emails.

- Too much email is a pointer to over-extension in other parts of life. Unsubscribe to newsletters you never read, get off useless lists, and involve yourself with fewer projects.

Good email habits and non-intrusive technology that lets people manipulate their digital information directly makes it pretty easy to handle about 500 emails a day, more than enough for most.
Link to this column

Tuesday, December 11, 2001

States proposed settlement The 9 rebel states rejecting the DoJ abdication have drawn up their own settlement proposal, requiring non-bundled Windows versions, restrictions on Microsoft middleware, open-sourcing IE, and several other monopoly-abuse busting provisions. Unsurprisingly, Microsoft is not impressed.

Given that Microsoft will never concede to this settlement and the DoJ is uninterested in the Law (or competition in the software industry), I'm not sure what the states are hoping for. Perhaps the plan is to force a renegotiated, and taking a hard position now will lead to a less useless settlement ultimately.

Monday, December 10, 2001

Letter to the DoJ I guess the DoJ has rolled over on consumers before, thus the Tunney Act. You all have 60 days to write to the anti-trust division and let them know what you think of the settlement. Remeber: be polite, and be constructive. Here's the email I sent today:

To: microsoft.atr@usdoj.gov
Subject: Microsoft Settlement

To: Renata B. Hesse
Antitrust Division
U.S. Department of Justice
601 D Street NW
Suite 1200
Washington, DC 20530-0001

Under the Tunney Act, I wish to comment on the Microsoft settlement's inadequacy in improving the competitive environment in the software industry. Some serious shortcomings relate to:

1) Middleware
The current language in Section H.3 states "Microsoft Middleware Product would be invoked solely for use in interoperating with a server maintained by Microsoft (outside the context of general Web browsing)" does nothing to limit the company's ability to tie customers and restrict competition in non Web-based networked services under .NET, as they fall "outside the context of general Web browsing". Microsoft has already begun abusing its desktop monopoly to tie customers int .NET revenue streams and set up a new monopoly over the network.

Part 2 of the same section states "that designated Non-Microsoft Middleware Product fails to implement a reasonable technical requirement..." essentially gives Microsoft a veto over any competitor's product. They can simply claim it doesn't meet their "technical requirements."

2) Interoperability
Under the definition of terms, "'Communications Protocol' means the set of rules for information exchange to accomplish predefined tasks between a Windows Operating System Product on a client computer and Windows 2000 Server or products marketed as its successors running on a server computer and connected via a local area network or a wide area network." This definition explicitly excludes the SMB/CIFS (Samba) protocol and all of the Microsoft RPC calls needed by any SMB/CIFS server to adequately interoperate with Windows 2000. Microsoft could claim these protocols are used by Windows 2000 server for remote administration and as such would not be required to be disclosed. The Samba team have written this up explicitly here: http://linuxtoday.com/news_story.php3?ltsn=2001-11-06-005-20-OP-MS

3) General veto on interoperability
In section J., the document specifically protects Microsoft from having to "document, disclose or license to third parties: (a) portions of APIs or Documentation or portions or layers of Communications Protocols the disclosure of which would compromise the security of anti-piracy, anti-virus, software licensing, digital rights management, encryption or authentication systems, including without limitation, keys, authorization tokens or enforcement criteria"

Since the .NET architecture being bundled into Windows essentially builds "anti-piracy, anti-virus, software licensing, digital rights management, and authentication systems" into all levels of the operating system, ANY API, documentation, or communication layer can fall into this category. This means that Microsoft never has to disclose any API by claiming it's part of a security or authorization system, giving them a complete veto over ALL disclosure.

4) Veto against Open Source
Substantial amounts of the software that runs the Internet is "Open Source", which means it's developed on a non-commercial basis by nonprofit groups and volunteers. Examples include Apache, GNU/Linux, Samba, etc. Under section J.2.c., Microsoft does not need to make ANY API available to groups that fail to meet "reasonable, objective standards established by Microsoft for certifying the authenticity and viability of its business." This explicitly gives them a veto over sharing any information with open source development projects as they are usually undertaken on a not-for-profit basis (and therefore would not be considered authentic, or viable businesses).

These concerns can be met in the following ways:

1) Middleware: Extend middleware interoperability with a Microsoft server to ALL contexts (both within general Web browsing as well as other networked services such as are those being included under .NET).

2) Interoperability: Require full disclosure of ALL protocols between client and Microsoft server (including remote administration calls)

3) General veto on interoperability: Require Microsoft to disclose APIs relating to "anti-piracy, anti-virus, software licensing, digital rights management, encryption, or authentication systems" to all.

4) Veto against Open Source: Forbid Microsoft from discriminating between for-profit and nonprofit groups in API disclosure.

Sincerely,

Zimran Ahmed
Link to this article

Friday, December 07, 2001

.NET is Microsoft's Internet Fork As Andrew Grygus points out, .NET is Microsoft's attempt to build new revenue streams now that the PC industry is a spent force (and will remain that way as Microsoft allows no desktop innovation from the hardware side, and is incapable of innovating itself). And in Microsoft fashion, this means building a monopoly and extracting monopolist rents. Sadly, the Internet is very far from being a monopoly, which is why .NET is essentially a fork that will split the Net into two camps, neither speaking to one another.

The Liberty Alliance is one telling instance of this split. It has Microsoft on one side, and Sun+Rest of World on the other. ZDNet, in typical clueless fashion, asks Microsoft to join the alliance, not understanding that this is antithetical to Redmond's corporate strategy.

I suppose it's good to see the States continuing anti-trust remedy negotiations and the Senate speaking to folks like Lessig about whether the DoJ settlement was too lenient. Perhaps we may see some sanity, perhaps not. Sooner or later, the government will figure out that monopoly power in the software world is about interoperability and bundling. Microsoft's insistence on "freedom to innovate" directly translates into "freedom destroy new markets that threaten our monopoly by bundling in products under proprietary standards." Any settlement that does not address this issue is worthless.

Thursday, December 06, 2001

McNealy and single-OS Scott McNealy has a good article on ZD-Net. He talks about Sun's single-OS strategy, and how that's different from IBM's abdication of platform control in favor of a true services model.

Wednesday, December 05, 2001

Sun and multi-OS companies A while ago I wrote about how multi-OS (IBM, Apple to some degree) and single-OS (Microsoft, Sun) companies were different. A reader wrote back with a dissenting opinion. He's a Sun employee, but these are his own thoughts.
Firstly, Apple is not a multi-OS company anymore than Microsoft or Sun is. Sure, Apple has both OSX and previous MacOS versions. But, Microsoft has Pocket PC, Win 95/98/Me, Win2000, and WinXP. Sun has Solaris 2.6/Solaris 7/Solaris 8.

Sun's positioning that they are a single OS company is really just a claim to focus. I remember a previous project (prior to working at Sun) where the choice was between Solaris, AIX, OS/390. Solaris was chosen, primarily because the IBM sales guys spent most of their time trying to discredit the other IBM platforms. The AIX guys spent their time convincing the client that OS/390 wasn't ready for primetime as a web based platform. The mainframe folks were discrediting the reliability of AIX. I've seen other equally bloody battles between the AIX and AS/400 guys.

I do believe that Sun's ability to have one hardware and software platform is a strength. I can use the same binaries on my Ultra10 that I use on the enterprise servers that my clients run. That's a good thing.

But Sun certainly isn't a single OS company as you define them. Not only are they very invested in Linux (http://www.sun.com/linux), but also produce software and hardware for a variety of other platforms. Including Windows (the JVM and iWS come to mind). Plus all of the Java support of non-traditional platforms like set-top boxes and phones. Sun even sells a Linux-based server (the Qube).

IBM is just a little more willing to market its Linux efforts. Is IBM's promotion a good thing for Linux? Sure. Would IBM be doing it if AIX was as successful as Solaris? Probably not. Does IBM sometimes have a lack of focus in its OS and hardware strategy? Yes.
Good points all. My response is that even though Microsoft has many OSes, their strategy is not around interoperation, it's around locking users into a platform and then extending that platform. All their OSes are part of a single-OS strategy, migrate people to and along Windows. Similarly, Sun's strategy focuses around moving users into and through Solaris.

Apple on the other hand, proprietary hardware and all, is such a marginal player that it can't afford to corral customers or pretend at being an entire infrastructure. Apple has to play nice with other vendors. Microsoft does not (and doesn't) and Sun would rather not also.

Internal squabbles between AIX, OS/390, AS/400 are just symptomatic of multiple divisions behaving like they are single-OS vendors, even though IBM's strategy has been to move towards "we'll use whatever OS is best for your needs and get it working with whatever it needs to work with". In the services business, lock-in is based on customer relationships and organization-specific expertise, not software. On the other hard, Sun's "everything would be great if you just used us for everything" is explicitly targeted towards moving businesses into a Sun environment and locking them in. IBM no longer controls any major platform, and they're OK with that. (And the Qube is a great little Internet appliance.)

Tuesday, December 04, 2001

Ximian's mixed source strategy Ximian's proprietary "Lucy" connector lets its GPL'd "Evolution" PIM speak with Microsoft Exchange servers. I don't quite understand why Ximian decided to GPL the PIM client and make the connector proprietary. A good PIM (or even a bad Outlook clone PIM) is not something that Free Software/Open Source would tend to create, making it a prime candidate for proprietary software. But an adaptor seems right up Open Source's alley, and would help commoditize the mail server under the PIM client, which is good for a proprietary PIM vendor. There is no similar benefit for a proprietary adaptor vendor for commodity OS PIMs. Making GNU/Linux desktops more viable in Microsoft server environment seems to be doing things the hard way, and keeping adaptors (which OSS tends to be OK at making) proprietary while GPLing bad GUI PIMs (which OSS is really bad at making) strikes me as being backword.

But it's easier now to have a GNU/Linux desktop in an all MS shop, which I'm sure will be good news for some. Personally, I'm bracing for the cataclysm when the Chicago Business School updates it's horrible IMP system to something that works with Outlook calendering. The school just suffered it's third Outlook based virus attack in about one month, so it's tieing itself to that God-forsaken platform more closely. *Sigh*

Monday, December 03, 2001

Microsoft will break TCP/IP Microsoft's horrible security record is a feature, not a bug. The use of raw sockets on Windows XP means the entire operating system will become a hotbed of viruses, which mostly propogate through the abomination that is Outlook on desktops. The Borg will blame TCP/IP and insist on its own, secure version of network protocol (which will only work with Windows, btw.) Moreover, this protocol stack will only support Microsoft media types, so goodbye Java, Quicktime, and Real, hello .NET. Microsoft's security guy going to work at the White House will make sure the law will back Redmond as it continues to put the Internet in its trunk.

(Who knows how true the above is, but it fits the fact and I wouldn't put it past them. It's difficult to see how anyone who knows anything about the tech. industry would be comfortable supporting Microsoft. The company has proven time and again that it's completely untrustworthy.)