Smalltalk Tidbits, Industry Rants

I love some of the rationales for not doing continuations on the JVM. Like this, from LtU:

There are a variety of reasons why we haven’t implemented continuations in the JVM. High on the list: continuations are costly to implement, and they might reek havoc with Java SE security model. These arguments are pragmatic and a tad unsatisfying. If a feature is really important, shouldn’t we just bite the bullet?

I just had that conversation with a friend before he left for JavaOne. He's frustrated by a number of things in Java, which all go back to the needs of the security model - his point being that it has less relevance for a server side application. I'm running this application on a Smalltalk server, where arbitrary code could be loaded in at any time. Here's the catch though - only two people have permissions on the system. So in order to mount such a code loading attack, one of the two of us would have to do it. Hmm - seems unlikely.

Anyway, that led me to Tim Bray's post:

This notion, that the Web GUI is insufficiently interactive and we need something richer, is widely held among developers and almost never among actual users of computers, and it’s entirely wrong. I can remember when people were forced to use compiled Windows and X11 applications, and most of them were extremely bad because it’s really hard to design a good interactive UI; when the Web came along, more or less everyone abandoned those UIs in favor of the Web, almost instantly and with shrieks of glee.

I'm calling BS on that. I'm not sure which user base Tim has met that generated that reaction, but either it was an oddball bunch, or the client applications they had to start with sucked. A lot. I deal with sales staff all the time, and they absolutely despise the web based sales tools that have been forced on them. Heck, most of them still use ACT! (or something similar) for their own use, and use the web system as little as they can. In customer visits, I get the same kind of reactions from people about the web apps they've had rammed down their throats.

Here's the thing: web apps aren't useless, and in the right place they solve problems very well. However, no one should be deluded into thinking that this is being done "for the user" - it's being done for IT, which hates having to manage client system deployments across a diverse range of hardware and software. That's not an unreasonable reaction by IT, but no one should confuse the desires of IT with the desires of end users.

This guy likes what he sees in Smalltalk, after coming to it from Java. He has a few qualms too, which is fair. I really liked this bit:

Here is a good article on the advantages of dynamic typing. I am becoming convinced that the “advantage” of static typing is that it provides minor babysitting services for bad programmers. Try writing a Java package that performs all sorts of calculations with int, double, float, and long data types and has to go back and forth between them…. the need to cast, convert, and develop special ways of handling “loss of precision” will drive you nuts.

Heh. I love that phrase: "minor babysitting services".

It's been another week of stats gathering, so it's time to look at the results. BottomFeeder downloads are up slightly, at 161 per day. The specifics:

Platform	BottomFeeder Downloads
Windows	508
Linux x86	124
Windows98/ME	111
Mac X	92
CE ARM	81
Update	53
Mac 8/9	51
HPUX	34
Solaris	26
AIX	17
Linux Sparc	16
Sources	8
SGI	2
CE x86	2
Linux PPC	2

Now a look at the HTML page accesses. Traffic levels stayed up last week, so there's been some gain due to the slashdotting. The tool details:

Tool	Percentage of Accesses
Mozilla	66.3%
Internet Explorer	23.2%
Other	4.7%
MSN Bot	3.4%
Everest/Vulcan	1.4%
Megite	1%

The Firefox share is staying pretty high too - it should be interesting to see whether the launch of IE7 has any impact on that. On to the RSS logs:

Tool	Percentage of Accesses
Mozilla	26.5%
BottomFeeder	16.3%
Net News Wire	9.3%
BlogLines	8.6%
Other	8.7%
Internet Explorer	5.1%
Safari RSS	4%
Google Feed Fetcher	3.4%
BlogSearch	2.3%
NewsGator	2%
SharpReader	2%
Planet Smalltalk	1.9%
RSS Bandit	1.5%
Java	1.3%
JetBrains	1.1%
Liferea	1%
News Fire	1%
MSN Bot	1%
Jakarta	1%
Everest/Vulcan	1%
Attensa	1%

Not a lot of change there - still lots of different tools being used.

Jeff Jarvis has a long screed up on books. After the first paragraph I was ready to rant - but his conclusions are no paeon to digital books. This is the part I was ready to pounce on:

The problems with books are many: They are frozen in time without the means of being updated and corrected. They have no link to related knowledge, debates, and sources. They create, at best, a one-way relationship with a reader. They try to teach readers but don’t teach authors. They tend to be too damned long because they have to be long enough to be books. As David Weinberger taught me, they limit how knowledge can be found because they have to sit on a shelf under one address; there’s only way way to get to it. They are expensive to produce. They depend on scarce shelf space. They depend on blockbuster economics. They can’t afford to serve the real mass of niches. They are subject to gatekeepers’ whims. They aren’t searchable. They aren’t linkable. They have no metadata. They carry no conversation. They are thrown out when there’s no space for them anymore. Print is where words go to die.

That's the opening though - it's meant to rile you up, so you'll read the rest. I had no idea that the number of new titles was declining, but Jeff has stats on that:

Publishing database Web site Bowker reported that there were more new book titles sold in Great Britain last year - 206,000 new titles, an increase of 28 percent - than in the United States -172,000 new titles, a decrease of 18 percent.

General adult fiction and children’s books both showed double-digit declines in new titles, Bowker found.

That gave me pause. Doc has written about how consolidation killed radio by making it universally bland; Dvorak has said the same about newspapers. I've generally liked the existence of big stores like Borders and B&N, simply because selection is better than it was at the tiny Waldenbooks that was the main bookstore where I grew up. However, we might be seeing the same thing in books that we see in radio and news: consolidation leading to a growing mass of same-ness.

I walked into the local Borders last night, in search of gift certificates. I should have taken a picture, because this point would be easier to illustrate that way. Right at the front, there's a table filled with new arrivals, and "The DaVinci Code" is still prominent there. That's not the weird part. The weird part is the next table, which is filled with books about "The DaVinci Code" - Which is a sign of the kind of growing blandness that killed radio, and is busily killing newspapers.

Thinking about this, I realized something about my own reading habits: I've been reading a lot of non-fiction lately, and virtually no fiction. That's a change for me from my younger days; I used to devour science fiction and military thrillers. Now? I'm boning up on the past. I think there's a connection there - as the kinds of books I used to like have all blended into one meaningless melange, I've discovered a rich vein of far better stuff.

I'm not sure that I'm completely down with Jeff on the need to move beyond the book to something that can be updated and annotated; there's value in reading the original thoughts of an author and having them fixed in time. There's also value in having the physical book; I find it's easier to read paper than it is to read screens - at least for long content. I suspect that the problems in publishing have more to do with what Doc Searls has said about radio than they have to do with the format.

If this kind of click fraud is as widespread as the article makes it sound, then the Google and Yahoo ad models are going to run into problems:

PandaLabs has detected a network of computers infected with the bot Clickbot.A, which is being used to defraud ‘pay per click’ systems, registering clicks automatically and providing lucrative returns for the creators. According to the data collected so far, the scam is exploiting a global network comprising more than 34,000 zombie computers (those infected by the bot).

The bots are controlled remotely through several Web servers. This allows the perpetrators to define, for example, the web pages on which the adverts are hosted or the maximum number of clicks from any one IP address in order not to arouse suspicions. Similarly, the number of clicks from the bot can be monitored as well as the computers online at any one time. The system used can evade fraud detection systems by sending click requests from different, unrelated IP addresses.

I understand that Google, Yahoo (etc) look for this kind of thing, and have systems in place to detect it. It sounds like those systems aren't working as well as they would like though - and if advertisers start to realize that, the rates that can be charged will drop. Given that all the air in Google's stock comes from advertising revenue... well, I'm visualizing a paper bag, just before the hand comes in. There could be a real explosion.

Via Doc Searls, I started reading here (Terry Heaton) and kept going down. From there, I ran across this story on CBS' Sunday night schedule. Paging down, I came across the list of shows, and that's when I had an epiphany: I have no idea which network most of the shows I watch are on. I rarely watch live TV; we cache stuff on the (yes, it's decadent) 4 DVR platforms we have. We then watch whatever matches our mood each night, without regard to the schedule.

The funny thing is, the riff Heaton has on the football schedule wreaking havoc on DVR recordings is something my wife has mentioned more than once: she hates it when sports whacks a show off schedule that way. What she hates even worse though is a practice NBC engages in - starting a show at :59 or at :01, which is specifically designed to screw with DVRs. Here's a tip for the programming geniuses behind that at NBC: It's not making me watch more ER. Rather the opposite, actually. This kind of thing is just anti-viewer, and it's not a way to win friends and influence people. Or, I'll just quote my wife on this practice: "ER just isn't good enough for that".

Ultimately, the schedule screwing is just going to have to stop. An increasing number of people are using DVRs of one sort or another, and they tend to be the people the producers and advertisers want. At the same time, the amount of available bandwidth for sending shows into the home is increasing all the time. I expect to see a couple of things to happen:

• Sports will move off the "main feed" for a network, so as to not interfere with other scheduling
• The kind of idiotic schedule games being played by NBC will come to an end

However, there's going to be an awful lot of whining and fusspotting between here and there. To get an idea of just how much whining, have a look at this from Heaton - if he's right (and I think he is), then the next few years of TV history are going to be extremely disruptive.

Cincom's booth at LWNW made it onto a news segment called "Webmania". Check it out here.

InformationWeek has an article about the rising tide of splogs out there. Splogs are spam blogs - and with so many free services (MS Spaces, BlogSpot, etc) out there, they are easy to put up. Heck, you can buy tools for it now. It's all pretty simple: you get one of the tools (or script one; automating the creation of blogs is not particularly hard), and then send your bot off to scrape content. Presto: Instant Splog. The purpose?

The people who create splogs--or, more accurately, the people who write the programs that create splogs--rarely intend for anyone to actually read their posts, which are often poorly written or even strings of nonsensical words. They're just building a giant clump of links that refer back to other sites, perhaps those that promote gambling or sell Viagra. When people click on those links, they increase the page rank of those sites on various search engines. Splog creators also sometimes include on their splogs ads that generate a small commission, usually a fraction of a dollar, for every click.

Here's one scenario: You want to test out a new programming language, so you run a blog search on it, hoping to find out about others' experiences with it. You end up at a site that looks like a blog--including a supposed blogger's name, photo, and archive of postings--but click on a posting, and you end up at a site advertising hard drive repair.

The search providers - from Google on down to blog specific sites like Technorati - are fighting this. The fact that I still get plenty of spam in my email inbox doesn't fill me with confidence about the end results though. Some of my search feeds are getting to be as useless as email because of this.

It's things like this, from SDTimes, that explain why I call Sun's OSS plans utterly insane. Last week, Jonathan Schwartz felt compelled to link here and try to explain how they can make money by not ever charging any. Well. In his long and rambling explanation, he Schwartz might have taken a stab at this:

In addition to the company’s reported losses, Sun announced an 8 percent drop in SPARC-based server sales, year-over-year. The company’s x86-based servers accounted for all of the company’s growth in this sector, and for 27 percent of the company’s overall server sales. Despite this, Sun saw a 12 percent increase in revenues in the United States, the first time that the company has seen growth in this country in two years.

Gee, what a shocker - after Java commoditized development tools, and after championing Solaris on x86 - and not charging for it - Sun's server sales dropped. Hmm. Doesn't take a rocket scientist to connect the dots there. Sadly, Schwartz can't find the dots, much less connect them.

Scoble on linky ads:

I don't mind this as much as I minded SmartTags when Microsoft was attempting to do them (before I was an employee I argued voiciferously against them, along with many other people in the community because we didn't want anyone to be able to use our own words for doing this style of advertising). Choosing to do it on your own blog only gets rid of most of my objection. I still don't like these kinds of ads, though, cause for someone who doesn't know the Web very well you can't tell these are ads at first.

Hmm. Seriously, how many people are there in the developed world who aren't familiar with the web? What's the actual possibility of confusion at this point?

Spotted in digg:

In an interview with SciFi Wire, actor James Callis, who plays Gaius Baltar, says that the new season of Battlestar Galactica which is currently filming and set to start in October, will be much darker, remorseless and relentless.

I had thought that the season two ending was pretty much all the way down, but I guess I'm wrong.

The wrong answer from management:

Companies are starting to ban Web access, block instant messaging services to squash discreet conversations among chatty co-workers and prohibit employees from watching sporting events on their computers.

"If you're watching video, you're probably not working," said Vimal Solanki, director of product marketing at McAfee Inc., a software vendor whose products to block Web access are selling briskly.

The correct answer from management:

Indeed, an employee for the New York City education department was recently fired for using the Web to read online news reports and visit travel sites while at work. He had been warned to stop.

That is not an isolated case. According to a survey conducted in 2005 by the ePolicy Institute and the American Management Association, 26 percent of employers have fired workers for misusing the Internet. A total of 526 companies responded to the survey.

The first example demonstrates a lack of work - by management. It's not really that hard to tell whether your employees are working or not; do they get their assigned tasks done? If they do, and they aren't violating any laws, why do you care what they are filling spare time with (yes, with some caveats :) )? If you do care, then you suck as a manager. Plain and simple.

The source of this - an article in the Chicago Tribune.

Avi weighs in on the topic Gilad Bracha brought up here, and references Tim Bray's post as well. He also brings in David Megginson. I posted on this here.

I just discovered Laura Ries' blog, after getting a copy of the book "The Origin of Brands". I'm still going through the older posts, which is why I'm hitting something from March. She highlighted a new store idea that WalMart is exploring - an attempt to move upscale:

In Texas, Wal-Mart opened a new store today to see if they can entice customers with a more glamorous layout, snazzy employee outfits and 1,500 new premium-priced goods. All this with the Wal-Mart brand name.

It’s true! The new store has hardwood floors, halogen lights, a sushi bar, and $500 bottles of wine. Employees will ditch aprons and don navy Polos and khakis.

She's skeptical about WalMart succeeding with that, and so am I - when you think WalMart, you simply don't think "high end". It reminds me of some of the confusion at ParcPlace-Digitalk and ObjectShare before the end back in the late 90's. After Java was released, management panicked, and decided that we had to have a Java toolset. I argued against that at the time, saying that it defocused us from our core product. That point was hammered home to me (and a sales manager who was stunned by this) on a customer visit. We went to an existing Smalltalk customer on Wall Street, and talked up our new Java tool. His response spoke volumes - "I may well be interested in java, but why would I buy it from you?"

The bottom line was, we were a Smalltalk company, and no one was going to take us as a "Java company" or as an "Object company". Which is where WalMart is going to fail with the attempt at a luxury position. No one sees them that way. Inexpensive clothing? WalMart. Expensive, marked up stuff that you buy to impress others? Nieman Marcus.

In short, the person who wants to boast about their conspicuous consumption isn't going to point at the new gadget in the kitchen and announce "I got that at WalMart!"

Michael Feathers goes public with the idea that "final" is just a bad idea:

Well... no.. not really. Here's the problem: When you use final pervasively, you make unit testing nearly impossible. Believe me, I know. I encounter enough teams who are in that situation. Have a class A which uses a class B that is final? Well, you'd better like B when you are testing A because it's coming along for the ride.

Extract an interface for B? Sure, if you own B. If you don't, you have to wrap it to supply a mock. And.. here's the kicker: if you are going to wrap B, well, what about security? All final does for a class is guarantee that the class itself won't be the point of access, but what about your wrapper? If you use the interface of the wrapper everywhere, again, because you want to test, the developers of B haven't made your software more secure, they've merely pushed the problem onto your plate: they've forced you to choose between testability and security. It's rather interesting to consider that perhaps we truly can have security, but only if we can't really be sure our software works.

The suggestion he has is the one that has been used in Smalltalk since the beginning: conventions that explain to developers which parts of a class' API might change in the future. What "final" does is attempt to lock the door before anyone's been in the house to look around. It assumes perfect knowledge about future usage, and that's a bad assumption.

Our friend Dave was kind enough to invite us out on his boat today - it was cloudy, but the win wasn't bad - at least on the way out. On the way back, the sun came out, and so did the wind. There was a a pretty rough chop out there on the way back - the wind took the canopy off! Anyway, I got a few decent pictures. Here's an automated lighthouse we passed:

Then there's this - a little hard to see, but there was a tiny little spot of land, and on it was a crane:

Here's the same thing, different angle:

This one should have showed more, but it's just a cheap phone camera. On the shore there are some of the expensive houses with their own piers:

Here's what it looked like coming back - although, you can't see the wind. It was brutal! Notice the clear sky though? Compare that to the first crane shot, which I took on the way out. In the space of a few hours, it went from cloudy to clear, and windy as heck. Still a great day to be on the bay though!

Via Don Box, I see that Ian Griffiths has a few issues with continuation based web applications. One issue he brings up is resource usage:

Sometimes the user just walks away. I might get part way through the purchasing sequence on a web site and then decide to stop. The web server never gets any positive indication that I closed the browser window. It merely stops hearing from me.

What does this mean in a world where I’m using continuations to help model user journeys as sequential code? It means that sometimes my functions just stop part way through without reaching the end.

On the plus side, it is predictable where this will happen: it can only occur at boundaries where I choose to construct a continuation and then relinquish control for now. But at every such boundary, I need to be aware that sometimes, the continuation will never execute.

This is very much not analogous to the function returning or throwing an exception. In the world of our chosen abstraction - that of sequential execution of a method - it looks like our thread has hung.

The problem with this is that a lot of the techniques we have learned for resource management stop working. Resource cleanup code may never execute because the function is abandoned mid-flow.

I don't know why this would be a problem, given a halfway decent GC system. The stale session will eventually time out, and take any lingering state down with it. I can see being worried about the memory required by this, in the face of a lot of users, but not in the basic mechanics. In a Smalltalk server - either VisualWorks or Squeak based - Ian's worry is a non-problem. Then he has an issue with thread affinity:

Traditionally, any particular invocation of a function runs on a single thread from start to finish. We’re not accustomed to mid-function thread switches, and it will render some hitherto safe practices unworkable. Using objects with thread affinity will become particularly hazardous, for example - we will need to be mindful of the potential switch points and make sure we never use such objects across such a boundary.

Actually, in this server, I've got requests that end up invoking new processes now, without continuations. They are managed with a class called Promise, but it's pretty simple - it's just a matter of delaying the waiting process and using a signal to inform the waiting process that results are ready. The pre-existing class makes that easy for me, but the code isn't hard to follow either way. Running multiple threads with shared objects is a comlex problem for any kind of application, whether continuations are involved or not.

Then he's got a long riff on how Continuations might break in the face of the Back button. Hmm - that was one of the problems that this approach was designed to solve. So far as I know, this simply isn't a problem in Seaside.

I get periodic questions about the state of WS* compliance in Cincom Smalltalk; here's the answer at present from our lead WS* developer, Tamara Kogan:

We have implemented:

1. SOAP 1.1
2. HTTP 1.1 RFC2616
3. XML 1.0
4. WSDL 1.1. (Currently supported only SOAP over HTTP binding)
5. UDDI 1.0
6. Compliant with Basic Profile 1.1 (There is an open problem with identifying SOAP faults. We still create SOAP fault message as it was specified in WSDL 1.1. We hope to address this in the winter release)

If that's not specific enough, send me an email and I'll see what I can find out for you.

Seems that Glenn Reynolds has the same opinion about the paranoid "ban the surfing" companies that I do. He's even more harsh than I was:

Sell your stock in companies with policies like this one. The management is obviously stupid, and the only employees likely to stay, long-term, in the face of this kind of a policy are those who can't get a job someplace else, someplace where the management is brighter than a bag of hammers.

Pretty much.

Macy's is going to put vending machines with iPods in them in the stores:

Macy's plans to install 180 iPod vending machines -- made by a San Francisco company -- nationwide by fall, its chief executive said Friday.

"This brings most-wanted merchandise into our stores in a unique new way," chief executive Terry J. Lundgren announced at the annual meeting.

That's a lot of quarters :)

Ian tried commenting here, but it apparently didn't work - that Javascript form may be more trouble than it's worth. Anyway, to the meat of his new post: he says that I missed his point, by assuming that by "resources" all he meant was memory. Based on his original post, that is how I took it. However, he explains further, and lists resources that GC doesn't typically handle: database connections, sockets, (etc, etc).

Here's why I didn't consider that - because all of that is at issue in any web app that maintains some kind of session state on the server. Whether you use continuations or cached data in some other form, you end up dealing with all of that (unless you re-open everything on each submit - which tends to get expensive). Typically, you use the finalization mechanisms of your language/library of choice to deal with that: the session times out, finalization kicks in, and you clean up. I fail to see how continuations make this more (or less) complex.

As to this on the back button amd continuations:

I'm not claiming that it's impossible to handle the back button this way. I'm just saying it's a bad idea, because you need to be able to write code that can tolerate being wound back and forth at the user's whim.

That's precisely the complexity that continuation based servers remove from your purview. In a "standard" we app, you tend to toss some kind of page key around, so that you can tell where you are (as opposed to where your app thinks it is on the server). How that's simpler, I have no idea. Having written a number of web apps in the standard way, I'm not prepared to call it simple.

Sony settled the class action lawsuit that came out of their rootkit fiasco - they still have a Texas suit to deal with though:

The agreement ends one chapter in a public relations disaster for the entertainment company, which must still contend with a lawsuit brought against it by the state of Texas for violation of state antispyware laws.

DRM software is bad PR, and - if done stupidly enough - might even be illegal. meanwhile, proving that they've learned nothing from this, Microsoft continues to plow forward with PVP-OPM. Here's a tip: the "Aero" interface is not cool enough to cover up that stench.

Troy mentioned that I should grab the latest version of the Javascript HTML editor we use here - it turns out that I was on a very old rev. I've updated to the latest, which just came out in May. They seem to be making progress on Safari support - not enough for me to make that the default for Safari though. I know some people have had trouble commenting with the editor - hopefully, this will be an improvement.

Dare Obasanjo points out that it's not a sign of the apocalypse when an MS site doesn't support Firefox (or Safari, etc):

Unlike Devanshu and Todd, I don't think there are sinister conspiracy theories for why two Microsoft products were released and ignored features of interest to the geek demographic. In every product release, you have a limited amount of resources and time in which to apply those resources to your next version. This means that you tend to focus on features that will provide the most bang for the buck and may ignore features that have limited appeal such as supporting a browser which is used by 8% of the market or a media subscription model is only used by 1% of internet users. I don't always agree with the practice of deciding on features based on market penetration statistics but I can understand when product teams make such decisions. I suspect that is more likely the cause of these omissions than some nefarious collusion between MTV and the Windows Media team or some plot to ensure IE's market dominance by having Windows Live services require only that browser.

Of course, it depends on who your market is. For some products, ignoring that 8% might be mostly irrelevant, while for others, it's going to be a killer. On the other hand, not supporting them will be an ongoing irritant. I was trying to use some award miles on USAirways. I normally use Firefox, so I logged in and went to the award pages. I got all the way to the end, credit card entered, seats selected, the whole thing. Submitted, and bam - tossed to the reservation page with no confirmation and no warning.

Hmm. I verified that no miles had been taken out, and tried again. Same result. Spent some time on the phone with an agent who couldn't figure out how to make the reservation for me (this has to go into the annals of bad service. She was apologetic, but could do nothing for me). Finally, I tried IE. At the point of submission, it slaps up some "please wait" page that must be IE specific.

Now, for the guy trying to cash in award miles, that's irritating, but it's no lost revenue for the airline. Had I been trying to buy tickets though? I would have been far more worried about the "did that go through" thing had each one involved a potential charge of a few hundred bucks.

So I agree with Dare - resources and inertia explain these things far better than some silly conspiracy theory. Having said that, the one taking the damage from this sort of thing is the vendor with the badly implemented site.

Cincom's President, Tom Nies, will be speaking at the AIT InfoSec conference at the UN:

18th Annual InfoSec and IT Infrastructure Conference & Exhibit

Theme: Moving Forward at Light Speed, the Latest InfoSec Threats, IT Alternatives and Solutions

Venue: The International Conference Center in United Nations Headquarters, New York City

Dates: June 27-28, 2006

Event Partners: IDG Pubs (InfoWorld, CSO), UN Global Alliance for ICT and Development

This Week’s Featured Presenter:

Thomas M. Nies, Founder, President, and Chief Executive Officer, Cincom Systems, Inc., is the longest-serving CEO in the Computer Industry. Tom will discuss, “I came. I saw. I simplified,” and explain how huge amounts of money are invested in InfoSec technology today. Tom will put into plain words the details needed to simplify information systems implementations, and will clarify how to maximize the impact of those investments.

Contact Steve Kayser for further details.

Wired:

An online real estate developer sues gamemaker Linden Lab for allegedly repossessing his cyberproperty.

If you're going to have a virtual world, maybe you need a virtual court system?

I added the same spam block system I'm using on the Wiki here, and turned trackbacks back on for some of the blogs as a test. Thus far, it seems to be working fairly well. If it gets overwhelming, I'll just turn them off again. If it stays sane, I'll enable them for all the blogs.

Fingers crossed...

Well. This story about how many of the early adopters actually "count" could have a twist - have a look at fakezilla:

FakeZilla.com traffic simulators are designed to send as many unique hits to your server as it can handle. Web page requests are routed through a massive list of anonymous proxy servers which can be defined by you. Counters and banners "see" these fake hits just as if a real user was browsing your site. When used with the Web Server Log extractor the fake hits and traffic appear 100% realistic- you can't tell the difference between FakeZilla traffic and real traffic!

Your venture capitalists can't tell, either - at least not until it's too late. It's simply amazing how many tools exist to create spam or fake traffic now.

There's an odd twist in the "iPod Nano Scratch" lawsuit - it looks like the lead plaintiff, one Jason Tomczak, never actually agreed to be a lead plaintiff. Via Digg, I ran across this open letter from him, where he says (in part - go read the whole thing):

The truth is that I never sought out nor did I ever hire David P. Meyer & Associates or Hagens Berman Sobol Shapiro to represent me in any case, much less the iPod Nano Class Action suit.

Later in the post he mentions that he sued the law firm in question, in order to try and clear his name. The results of that demonstrate that even well intentioned laws can backfire. The law firm filed an Anti-SLAPP motion against him, using California's protections against SLAPP suits. Seems completely ridiculous to me, but dealing with that will chew up time and money - both of which are in greater supply for the law firm.

I googled Jason's name, and ran across something very curious - on the leading pages that turn up, he's not quoted. The lawfirm is, and his name comes up, but there are no direct quotes. That seemed fishy to me, and also seemed fishy to Couthouse News. It looks a lot like the lawyers who filed the suit against Apple went for an easy payday against a large firm (on the old "cheaper to settle than fight" theory of operations) - and they didn't even take the time to find a real plaintiff.

In the meantime, I'm going to try and contact Tomczak and get his side of this directly. There doesn't seem to be a working email address, but the open letter does have a snail mail address. I'll try that, and I'll post on this again when I know more.

Update: Slashdot picks up the story.

I had a checkup this morning, and after seeing the cholesterol numbers, I'm motivated to go out for a jog now. I'll be back online this afternoon

Who knew that Hacky Sack was a professional sport with a world championship? RocketBoom has the video.

Spotted in Scripting News

I love and admire Mike and Richard, and I'm glad they're welcomed by the owners of Web 2.0, but until they put out a welcome mat for everyone else, I'm going to keep looking to the future, because I think that kind of exclusivity belongs in the past.

And Mike, if you wanted to get rid of the problem, one call to O'Reilly or Battelle right now would probably take care of it. And mention it to Kevin Werbach as well. That you and so many others quietly acquiesce allows the exclusivity to continue. Until then, I'm going to keep looking for a route-around, and some day, hopefully soon, we'll find it.

Here, let me translate: "Wahhhhh - they won't invite me to their conferences. Wahhhhhhhh"

Spam has made email an unreliable mechanism. Witness this case of a lost bid:

There it was in the e-mail spam filter, along with offers to invigorate both your bank account and your sex life: an offer to save the Cobb County schools $250,000. But this message was for real.

School officials are blaming an overeager junk-mail filter for capturing and killing a Kennesaw businessman's bid to provide telephone services to the system. It seems the part of the filter that watches for pornographic material was offended by the use of terms such as "long distance."

I have to check my junk folder (both the corporate server-side one, and the local client-side one) regularly - otherwise, I end up missing quite a bit. This has tragedy of the commons written all over it.

In News that should shock no one, Microsoft's Steve Ballmer is dampening expectations about the release date for Vista:

The operating system was due to be launched this year but in March the company said it wouldn't get broad release until January 2007. Ballmer said Wednesday that the planned January launch may slip further based on feedback from a beta release program and the product road-maps of hardware vendors.

That's actually a smart move, getting out in front of a potential delay like that. With feature driven releases, timing is always hard. We do timeboxed releases of Cincom Smalltalk, but the consequence for that is that the delivery of any given feature is always up in the air. You get predictable features, or a predictable date - you don't get both. All too often in this business, you get neither :/

Terry Heaton, via Doc Searls:

Because here's the deal. The tools available to everyday people that are turning the media world on its head are also available to professional organizations. You don't have to approach everything with a $100,000 solution when $10,000 will do just fine. If aggregation is where its at (and I believe that it is), then build aggregators. Let other people be the content creators and move yourself to the edge. Not only is it fun there, but that's where the profitability is going to be downstream.

Isn't that pretty much what Amanda Congdon and RocketBoom are doing right now? Over time, I expect that most news operations will look a lot more like movie productions - you'll have a small number of permanent employees who manage a wide range of stringers, whose expertise varies and is available as needed.

Whether Google like it or not, being in the media business (and Google News is a media outlet) requires some transparency on their part. Steve Rubel points to the problem:

Google is getting smacked for removing conservative e-zines and blogs from Google News. To date, Google has done a poor job of explaining why certain blogs are enshrined and others aren't. They need to publish some standards around who they are willing to include and why.

What content they are pulling isn't nearly so relevant as why it's being pulled. Like a newspaper, they have no requirement to publish anything they don't want to publish; it's their site, and they can do what they want. If, on the other hand, they want to retain actual credibility, they need to explain what standards they use for including (and not including) content.

As if "final" isn't bad enough, some people would like to make it the default:

I’ll say it one more time: final should be the default. Java’s mistake was not that it allowed classes to be final. It was making final a keyword you had to explicitly request rather than making finality the default and adding a subclassable keyword to change the default for those few classes that genuinely need to be nonfinal. The lack of finality has created a huge, brittle, dangerously breakable infrastructure in the world of Java class libraries.

Umm, sure. Because every library I've ever used was designed specifically with my needs in mind, and I've never, ever had to extend anything.

*Cough*

That must have been something coming up there. I guess the BS meter just went to infinity, and I couldn't take it anymore. He goes on:

One final point: final is the safe, conservative choice. Should you mark a class or method final, and later discover a need to subclass/override it, you can remove the finality without breaking anyone’s code. You cannot go the other way. Once you’ve published a class that’s non-final you have to consider the possibility that someone, somewhere is subclassing it. Marking it final now risks breaking people’s running code and working systems.

It's safe, just like a straitjacket. With all possible movement impossible, how much can you do? It's not about breaking someone else's code anyway. When I get a library from a vendor, I don't expect perfection - I expect a best effort, given the knowledge the engineers had at the time. I fully expect to have to subclass in some places, to delegate in others, and to override methods in subclasses in some cases. Heck, I'm using smalltalk, so my amp goes to 11 over here - I may even change code in existing classes in the library. I'm sure that concept will make his head explode, but my goal is to empower the developer - not to break his wrists because he folded the napkins wrong.

I guess it's been awhile since Nick Carr got some blogosphere love - today, he's calling Wikipedia "dead":

Wikipedia, the encyclopedia that "anyone can edit," was a nice experiment in the "democratization" of publishing, but it didn't quite work out. Wikipedia is dead. It died the way the pure products of idealism always do, slowly and quietly and largely in secret, through the corrosive process of compromise.

Right. Apparently, restricted editing rights on some of the more controversial pages signals the end times. Here's his point, such as it is:

The end came last Friday. That's when Wikipedia's founder, Jimmy Wales, proposed "that we eliminate the requirement that semi-protected articles have to announce themselves as such to the general public." The "general public," you see, is now an entity separate and distinct from those who actually control the creation of Wikipedia. As Vaughan-Nichols says, "And the difference between Wikipedia and a conventionally edited publication is what exactly?"

Sure Nick, sure. And blogs that moderate comments are dishonest, too. Maybe I should set up a chart, so I could plot Carr's needs for attention and see if there's a pattern.

If you are looking for an overview of the state of grid computing, then head on over to Tim Bray's place.

Fnordistan explains how to package up a Cincom Smalltalk application into a runtime - including the creation of a Windows executable. It's easy to follow, and has all the steps you'll need to follow. Highly recommended.

Colin Putney is working to make Monticello a cross dialect version control system for Smalltalk:

However, there is a real need for tools to make cross-dialect development easier, and versioning is an important component of that. After doing a few demos, I had volunteers to maintain VisualAge and Dolphin ports. The VisualWorks folks all seem pretty busy, but I'm sure somebody will step up when MC2 gets to production quality.

I'm not surprised that Randy Johnson is having a tough spring - he's an old power pitcher, and - other than Nolan Ryan and Roger Clemens, power pitchers tend to fall fast as they age. I thought it was a mistake to get Johnson when they got him, and last night's performance is just giving me more evidence:

But on Wednesday, he allowed at least five runs for the sixth time in eight starts before improving after Jason Varitek's single put Boston ahead 5-4 with no outs in the third. Johnson struck out the next three batters and retired nine of the last 10 he faced before leaving after five innings.

"I hit my spot better later in the game,'' Johnson said. "It's funny to say this, but what I'm going through, you take anything as a positive. The last three innings were a positive. How much worse can it get?"

Despite his solid finish, his ERA rose from 5.62 to 5.89 while he allowed nine hits and two walks with eight strikeouts.

Here's a tip for Steinbrenner: If the pitcher is pushing 40, then don't hire him. Develop the talent out of the farm system, for gosh sakes.

The way they ended "Lost" last night wasn't as mind boggling as the ending to Battlestar Galactica's season was, but it does make for an interesting beginning to the upcoming third season. As SciFi Wire said:

Producers of ABC's hit series Lost, which aired its surprising second-season finale on May 24, told SCI FI Wire that it sets up the upcoming third season, which will focus more on romance -- and on the mysterious Others. "The Others are an important part of season three, and there's a lot of mysteries and a lot of questions about the Others that the audience is going to be curious about going into season three," said executive producer Carlton Cuse in an interview. "And those are the things we're going to explore."

Follow that link only if you don't care about spoilers, or have seen the last episode. With the wrap on the regular season at hand, I'm waiting for July, when SG-1, SG-A, and BSG start up again.

Update: As James points out in the comments, BSG isn't back until the fall.

SharkTank relates a humorous support incident, where a caller has to lie in order to get support. It's posted as a funny anecdote, but it's actually an illustration of what happens when you ship support to a third party - to manage it, you apply a script and rules. That keeps costs low, but it torques off customers.

The amazing thing is this: everyone mouths the idea that it's easier to sell an existing customer than a new prospect, but then thye go out and act as if the goal is to treat the existing customer like a musroom - in the dark, and covered with manure. As word of mouth PR spreads, leaving existing customers with a bad taste in their mouth is just going to get stupider and stupider.

Nintendo has announced that the Wii will not cost more than 25,000 yen (about $225 USD). Here's the quote:

The player won't cost more than 25,000 yen, said Yoshihiro Mori, Nintendo's senior managing director, at a news conference in Osaka. The company expects to ship 6 million Wii consoles this fiscal year, and about 17 million software games for the device.

Given that, I'm going to make a prediction: The US price will start at $199. That puts it down into the psychological "impulse buy" space for an awful lot of people in the US. I guess we'll have to wait to see if I'm right.

What Wired is calling Crowdsourcing is an example of what Glenn Reynolds wrote about in "An Army of Davids" - for any given field, there are underused experts out there. This is going to be a far bigger driver for business change than offshoring is - why go through the hassle of dealing with people half a world away when you can grab the spare cycles of people who are no more than 3 timezones away? The example here has to do with photo licensing, but it applies to things far beyond that:

iStockphoto, which grew out of a free image-sharing exchange used by a group of graphic designers, had undercut Harmel by more than 99 percent. How? By creating a marketplace for the work of amateur photographers – homemakers, students, engineers, dancers. There are now about 22,000 contributors to the site, which charges between $1 and $5 per basic image. (Very large, high-resolution pictures can cost up to $40.) Unlike professionals, iStockers don’t need to clear $130,000 a year from their photos just to break even; an extra $130 does just fine. “I negotiate my rate all the time,” Harmel says. “But how can I compete with a dollar?”

He can’t, of course. For Harmel, the harsh economics lesson was clear: The product Harmel offers is no longer scarce. Professional-grade cameras now cost less than $1,000. With a computer and a copy of Photoshop, even entry-level enthusiasts can create photographs rivaling those by professionals like Harmel. Add the Internet and powerful search technology, and sharing these images with the world becomes simple.

This is going to hit business in any task that isn't completely core, and can be shopped out. Over the next few years, the shops that thought they were cutting edge by moving operations abroad are going to learn that they are already behind the curve.

Well, this is nice. I had to take my daughter to an activity, and there's a Barnes & Noble nearby. The WiFi isn't free, but it's not expensive either. It's pretty nice to be online, with a Mocha, with the history books right down the steps...

Here's a Negative PR Event bearing down on O'Reilly - they think they own the term "Web 2.0":

One of these events - the upcoming Web 2.0 half-day conference is the target of a cease and desist letter (below) from the legal team of O’Reilly publishers. Basically O’Reilly are claiming to have applied for a trademark for the term “Web 2.0″ and therefore IT@Cork can’t use the term for its conference. Apparently use of the term “Web 2.0″ is a “flagrant violation” of their trademark rights!

So how long do you think it will be before Tim O'Reilly notices the "boy, what tools those clowns at O'Reilly are" flareup over this? Will he call off the dogs quickly, or will the negativity have to rise to Warner Kremer Paino levels first?

Who knew that Spock could get down?

If you're looking for high level information on our Web Services implementation(PDF), check out the new whitepaper we've put out. Those of you looking for information on ObjectStudio 8(PDF) should look at the new whitepaper for that.

The RIAA is just amazing. First, they say that suing 12 year olds and grandmothers is just peachy:

When asked if they regret suing people like 12-year-old girls and grandmothers, Sherman says no, and that they're "feeling pretty good", then goes on with some lines that pay lip service to the idea that they're interested in coming up with new business models, rather than just using litigation as the cornerstone of their strategy.

That's bad enough, but then they whipped out what techdirt refers to as an attempted jedi mind trick:

The most egregious comment, though, comes from Bainwol, who says "nobody" has any problem with DRM and copy protection. While consumers might not know what DRM is, they know when they songs they've purchased won't play on their new MP3 player, because it's not compatible, or when they can't burn a CD to their computer because a record label thinks they're a criminal, or when the copy protection on a CD opens their computer up to hackers.

Sheesh. I think Tony Soprano would be ashamed to be in the same room with these clowns.

Looks like Chris Pirillo has found more things to dislike in the new Outlook than things to like.

I think the problem is simple - developers don't know when to stop, and management doesn't want them to stop. Office is a huge chunk of MS' revenues, and there's a corporate need to drive upgrades. That's why management wants churn. However, take a look at what is, for the most part, a mail client. Is there really that much more it needs to do? Is the complexity inherent in making it a combined PIM, email client, and RSS reader worth it?

Based on what most users of Outlook seem to say, I don't think so. It's too big, too bloated, and too unfocused on any of the things it does. What I think MS really needs is 3 smaller, lighter tools here that can communicate in some fashion. That's not going to happen though - instead, the suckage will just increase.

The pixels on my monitor nearly revolted after I browsed this page:

Operate All Revenue Generating Channels in a Web 2.0 Architecture by 2008. Enterprise architects must act as catalysts that speed the formation of unified business technology strategies and their execution. The enterprise architecture process must shift gears from limiting complexity by limiting choices to accelerating innovation and execution by coordinating complexity through unified business and IT strategy, decentralized execution and loose coupling among all related stakeholder disciplines.

That's, umm, something from a Gartner analyst at the recent ITxpo in San Francisco. I wonder if the participants were aware of just how much oxygen was being wasted. With luck, the room had some plants so that something benefited.

Hat tip Stefan Tilkov.

Update: Heh. In the IRC channel, one of the regulars had this comment: "It says that the enterprise architect shouldn't just limit choices, but decide for others too. Lend your business to the architect, so he can ruin it more quickly."

Looks like O'Reilly and CMP recognize a PR problem when they see one coming down the pike - they've called the dogs off of the IT@Cork conference. They still insist that they own the "Web 2.0" moniker though, and state that it's like the term "LinuxWorld" - a stock use of a trademark:

O'Reilly and CMP co-produce the Web 2.0 conference. "Web 2.0" was coined when we were brainstorming the concept for the first conference in 2003. As noted in the letter to IT@Cork (sent from CMP's attorney, but with our knowledge and agreement), "CMP has a pending application for registration of Web 2.0 as a service mark, for arranging and conducting live events, namely trade shows, expositions, business conferences and educational conferences in various fields of computers and information technology." To protect the brand we've established with our two Web 2.0 Conferences, we're taking steps to register "Web 2.0" as our service mark, for conferences. It's a pretty standard business practice. Just as O'Reilly couldn't decide to launch a LinuxWorld conference, other event producers can't use "Web 2.0 Conference," the name of our event. In this case, the problem is that it@cork's conference title includes our service mark "Web 2.0," which the law says we must take "reasonable steps" to protect. We've also contacted another group that has announced a "Web 2.0 Conference" in Washington, DC this September.

Here's the problem - the term "Web 2.0" is way, way too widespread for them to claim ownership. Here's a Google search on it - see what I mean? The law may well be on their side, but common sense very much isn't. If they think they're being harmed when others use that term, then they need an appointment with a cluestick.

Seaside is making its way to Dolphin - there's an early port available here. Seaside now works on VW, Squeak, and Dolphin - if the Instantiations folks step up like the Object-Arts folks did, it could work on VA too.

We've posted another white paper - this one is on the Security Libraries that ship with Cincom Smalltalk. Download the PDF here.

Thank goodness we have smart folks like Nick Gall over at Gartner - otherwise, a term many of us have become familiar with - REST - wouldn't be pushed aside for an enterprisey acronym like WOA. Without the extra buzzwords, how would enterprisey architects, and the analysts who support them, ever manage to validate their existence?

There's video of some of the sessions online here. The keynotes from Doc Searls and Dan Gillmor are there, as well as the session with Dave Weinberger. I haven't sat through them (I was there when they happened), but the quality from the bits I sampled looked pretty good.

Here's an interesting paper that comes from No Fluff, Just Stuff (PDF) - I particularly agree with this summation towards the end:

I predict that we’ll see the increasingly wide adoption of dynamic languages, metaprogramming, and agile design and development practices over the next few years. In spite of many naysayers, momentum seems to be building in this direction.

I don’t think it will stop with Ruby, Python, or any of the other new old languages that are gaining popularity. Although those languages borrow extensively from their progenitors, they stop short in some other ways. I love programming in Ruby, but occasionally I find myself needing some of the features of Smalltalk or Lisp that Ruby doesn’t have -- true macros, for instance, or the ability to easily pass multiple blocks to a single method (with appropriate cues as to their distinct roles). And don’t get the idea that I’m an old Smalltalk or Lisp programmer! I come from a C, C++, and Java background. But I’ve recently begun to understand some of the subtle strengths of languages that I used to think were weird.

You have nothing to lose but your lower levels of productivity :)

The mini seems to have died. I turned to put some music on, and it was locked up. Couldn't get in via ssh, so I turned it off. Now, it won't boot. Won't boot from CD, even. Guess it's off to CompUSA. Sigh.

"XMen: Last Stand" was ok, but nothing special. It took an awfully long time to get going. By 30 minutes in, I was waiting for the ending which - while it was entertaining - went down pretty much the way I expected it to. It also felt choppy at times, as if large pieces of characterization had been left on the cutting room floor. It's a decent enough "popcorn" movie, but don't go expecting to be awed.

One thing I will say - stay in the theater all the way to the end of the credits. There's a scene that's shown then, and if you get up as the credits roll - no final tidbit for you :)

It's that time again - Saturday morning. I'm off to Annapolis with my daughter's girl scout troop today, but I've got the weekly numbers first. BottomFeeder downloads proceeded at a rate of 144 per day - down from last week, but in the normal range. I think the bump from the slashdot effect has disappeared:

Platform	BottomFeeder Downloads
Windows	487
Linux x86	115
Mac X	87
CE ARM	80
Mac 8/9	57
Windows98/ME	48
Update	42
HPUX	26
Linux Sparc	22
Solaris	18
AIX	17
Sources	11

On to the HTML page accesses - towards the end of the week, the total pageviews tailed off, while the number of unique pageviews (by IP address stayed about the same. More RSS? Fewer scripts trying to leave spam? The tools:

Tool	Percentage of Accesses
Mozilla	66.7%
Internet Explorer	24.1%
Other	3.2%
MSN Bot	2.6%
Megite	1.3%
Everest/Vulcan	1.1%
Opera	1%

Interesting that Safari doesn't show up at all in the HTML accesses, but in the RSS accesses, Mac tools figure so prominently. Speaking of which, those numbers:

Tool	Percentage of Accesses
Mozilla	26.5%
BottomFeeder	17.8%
BlogLines	8.7%
Net News Wire	8.2%
Internet Explorer	7.6%
Other	8.9%
Safari RSS	4.1%
Google Feed Fetcher	3.5%
BlogSearch	2.2%
NewsGator	2.1%
SharpReader	1.8%
Planet Smalltalk	1.4%
Java	1.1%
JetBrains	1.1%
RSS Bandit	1%
Liferea	1%
Everest/Vulcan	1%
News Fire	1%
Jakarta	1%

And I'll be back later, after the day trip

Bloggers won a round against Apple - Wired reports that Apple could not compel PowerPage to reveal their sources for a story they posted on (then upcoming) Apple plans. To boil it down, the court said that it was journalism in action, and that Apple couldn't stop bloggers any more than they could stop mainstream publications:

A California appeals court has smacked down Apple's legal assault on bloggers and their sources, finding that the company's efforts to subpoena e-mail received by the publishers of Apple Insider and PowerPage.org runs contrary to federal law, California's reporter's shield law, and the state Constitution.

This is good, but it's not the full extent of rights that we ought to expect, at least here in the US. Right now, if you post on politics, and you use AdSense (or accept any other type of ads), you could conceivably run afoul of various campaign finance laws - which have the effect of muzzling what you might want to say. The big media needn't worry about that; they have their press exception. The thing I've always wondered about that is simply this: why does a big outfit like GE (owners of NBC) get more rights than I do?

Anyway, this Apple decision is a good thing - but it's cause for a satisfaction, not joy. At least IMHO.

It's nearly June, and the 2006 hurricane season (for those of of us in the eastern US) is upon us. I thought I'd take a look at the historical trends for hurricanes, and there's a lot of data out there. This site has a nifty table of frequencies (up to and including 2004) - and it illustrates something interestesting: notice how frequency had a local peak in the 1941-1950 interval, and then started to drop? That coincides with the end of WWII, and the rise in prosperity in the US. It also coincides with the spread of inexpensive air conditioning.

What does that mean? It means that people started to move south in large numbers. In particular, coastal development ramped up - along both the US east coast and along the gulf coast. It almost certainly seemed safer - the table tells the story. According to that data, the annual hurricane frequency has been below average since about 1950.

Mind you, the data only goes back to 1851, and that's a very short time in climate terms - the "average" for such a short time span could well be meaningless. In behavioral terms though, it's not. The building boom along the coasts coincided with a period of lower activity, and people didn't remember that really bad storms (1935, for instance) were not just possible, but highly probable. We tend to completely ignore things that we don't have personal experience with.

So, have a look here; while there were two absolutely awful hurricanes last year, the raw numbers were 15 storms of category 1 or higher (storms get names if they hit tropical level - category one is much stringer). In 2005, 15 storms hit category 1 or higher. And out of that 15, you have to drop the ones that never got to the US mainland in order to fithem into that table. That data is further down in the second tabll, and it was 5 storms. So from 2001-2005, that gives us 14 storms that got to level one or above, and also hit the US mainland.

Which means that we are back into the kind of active period that we last saw (again, here in the US) back in the 1940s - before the current buildup on the coasts happened. What does all that mean? Darned if I know - in looking at the data since 1851, it looks like we've been in a relatively quiet period since 1950. As I said above though, this is such a short interval in climate terms that it may be meaningless. The only real conclusion I can come to is this: it was probably a mistake to build as much as we have near the coasts. I'm guessing that insurance rates will reflect that pretty quick.

If you want to track these storms yourself (I do, since my parents live on the east coast of Florida), then you should subscribe to the NOAA tropical storm feed.

We headed down to Annapolis for a girl scout outing today - it was beautiful weather, so being a chaperone was a pleasure. We started the day with a walking tour of the downtown area. My wife had set up a guided tour, which was very nice. We went by the harbor, the state capital, and into the kitchen of one of the historic buildings - set up as it would have been back then. The capitol is quite nice - it's the oldest, in continuous use capitol in the US. Here's a picture I took, looking up the street towards the capitol dome:

The harbor is quite nice too - you can't see it in the low res camaera phone shot I took, but the Bay Bridge is off in the distance here:

It was cloudy for a bit while we were there, but it cleared off and got quite warm. We headed up the street from here to the Paca house. William Paca was one of four Marylanders who signed the Declaration of Independence; he had a house with a view of the Severn River built in the 1760s. It's no longer got that view; the Naval Academy is in the way (along with a bunch of filled land). Here's a view of the Paca house from the garden:

And a a view from the house, looking down at the "summer house" on the far end of the garden:

The house has an interesting history. It was privately owned until the turn of the 20th century, when a hotel company took it over. They left the house mostly alone (although they modified the bedrooms to add bathrooms, so they could be sold as suites). The garden was paved over and the bulk of the hotel was built there. It stayed that way until the early 60's - at which point the entire structure was nearly levelled for a parking garage. Fortunately, a foundation managed to raise money and buy the house and land, and then spent a number of years in restoration. It's a beautiful spot now - well worth visiting if you happen to be in Annapolis.