1,431 karma · joined August 25, 2009
What if the user wants to have more than one feed? Or wants, sometime down the road, to have routable resources that are not feeds?
Wouldn't it be better to say that a user name is a user name and that a default feed name can be automatically built given just the user name?
Using DNS to identify users is unwise, in my opinion, because it means that people won't own their own on-line identities -- they'll have to rent them and for real money, too. And if some users are assigned a sub-domain on a shared domain, their identity won't be portable.
I think it is worth doing a little extra work to make a user name system that doesn't have those problems.
I believe you missed a "point of importance and ultimate goal" of the open source movement: to preserve and enhance the market for proprietary software.
On that goal, free software and open source are diametrically opposed.
I think there's only one conclusion:
The web itself is in decline.
Incredible, I know, but this is science.
That's part of why you don't program in Common Lisp all the time.
Bodies of law (e.g., the Berkeley municipal code, the CA state code, or the federal code) are organized into various titles, divisions, chapters, sections, subsections, subsubsections and so forth. Each body of law is a kind of multi-volume publication, structured that way.
Legislation - the acts passed by legislative bodies to modify the law -- are expressed as patches (diffs). Legislation says things like "In title X, division Y, chapter Z, section M, subsection N, strike the words "blah blah blah" and insert the words "blather blather blather". That's a patch. Each act of the legislature (that changes law in this way) is a patch. The law is version controlled. (In fact, there's an important role in government for the archivists who receive these patches, apply them, and keep the official record of the "trunk" branch of law up-to-date.)
Beyond that: records of these patches are kept meticulously. If you have the record at hand, and you are reading some sentence somewhere in the law -- the record can show you all the past "diffs" that led up to the sentence as it currently stands.
It gets even better. Legislative bodies keep archived records of their discussions and proceedings in the course of passing law. Think of these as "comments on a commit". When you are looking at a passage in the law, and want to study its commit record -- it's all there for you. In court, this is called the "legislative history". The legislative history is not itself part of the law but, as with commit comments, courts sometimes find that commentary helpful for understanding "the code".
Fancy (mostly but not always for-pay) interfaces to law on-line actually let you explore these diffs and comments pretty directly as you read the (legal) code.
I'm not sure what the state of the art is for software in support of legislators and lobbyists -- that is to say, for the "programmers" that write and submit "patches" to law.
So my suggestion is that you do competitive research on all the on-line viewing and authoring tools, look at what capabilities they offer that seem important, develop free software alternatives that use version control concepts to help expose and make a nice interface to the version controlled structure that is already there -- and then (if you are trying to build a business) -- work on a plan to populate your software with data (not easy or cheap) and sell services atop that.
In addition to the moral reasons you should make it a free software project, a good business reason is because the market is already crowded with expensive proprietary solutions. So you can make a killing by shrinking an N-billion dollar market to be much smaller -- but much more owned by you (the shrink-the-market "Red Hat trick" for unix-on-the-server that, per the mythology, largely killed Sun and raised Red Hat).
The note describes memoization as using a "cache". That's a problem because memoization and caching are both useful, but they are different things.
A memo (which it appears is implemented by the Python library you mention) stores the results of earlier computations for re-use -- just as you've described.
A cache does the same thing except that if a cache becomes over-full, some earlier cached results may be recomputed (because they were dropped from the cache).
There are two critical differences:
Minimum space requirement for a memo: O(N x S) where N is the number of unique computations to be performed where S is average size of parameters and results.
Minimum space requirement for a cache: arbitrary, although performance varies with size.
Number of repeated computations with a memo: 0
Number of repeated computations with a cache: at least N where N is the total number of unique computations, and as many as T where T is the total number of computations with duplications.
A cache is useful in contrast to a memo when N*S is uncomfortably large, but "most" of the repeated computations are members of a set of size M where MxS' << NxS where S' is the average size of params and results in the M-set.
A memo is useful where NxS is conveniently small, or where it is unacceptable to repeat any computation for some reason.
As a point of ... trivia? ... Eben Moglen, the attorney that created the GPL along with Richard Stallman, directs and is a founder of SFLC.
Here's a wish-list (vis a vis commercial web sites):
Please...
Always have an "about" link that tells me the name of the company, its corporate form, and who non-anonymously stands up to represent the owners. I want to to know, before we do business, that you are legitimate and not hiding behind a veil of anonymity.
Please include a "contact" page that includes a regular ol' email link, a valid street address, and if at all practical, a phone number. I want to know, before we do business, how far I have to walk to knock on your door in case there's a problem.
Please have a page that explains, concisely yet substantially, what the new site/service does. I do not mean "Do you have problem X? We solve X!". I mean a non-salesy, short guide: "If you sign up you get Y. Y performs function Z."
And as long as I'm wishing for candy and nuts: please have a page for technical types explaining in broad terms how the service functions. Now, I understand that this can be sensitive. You don't want to give away the recipe for your "secret sauce". Still, you can hopefully tell me enough technical info to give me a realistic model of what it is that you propose I use.
For me, when I look at a new or new-to-me service, those are the questions I'm hoping to find answered in the first couple of minutes of looking. When a few of them aren't answered by the site, I tend to give up on the site after just a few minutes.
It doesn't stop there, though. Some software vendors also like these restrictions. With treacherous computing, vendors can make themselves the exclusive providers of software and enjoy monopolistic pricing. In contrast, consider that the cost of a complete operating system for a non-broken computer starts with many decent options at a price of $0.
Treacherous computing also makes it easier for software vendors to monitor their users in ways that users can't control and might not even notice. There is no way, for example, to tell some devices not to "call home" and betray a users privacy to their vendors.
So it's a convergence of reasons, not just the "entertainment industry" (I guess you mean a few big companies, not everyone who sells entertainment) behind the thing.
The free software movement seeks to educate all people about software freedom and to encourage them to value and protect software freedom.
When someone buying a product "doesn't care" about these restrictions, especially if the buyer doesn't understand what they are giving up, that's a problem the FSF wants to solve.
So, yes, by all means: if you don't want those restrictions, don't buy those products. The Free Software Foundation would like to help more people come to that same conclusion, rather than naively buying into a software world where users no longer have much control over their own computers.
The "theatrics" you refer to -- picking alternative, damning names for treacherous computing devices -- are very much present in this article:
Instead of "Secure Boot," the article says, if the vendors implement it in a way that deprives users of control, we should call it "Restricted Boot". Even the reason for the alternative name is similar to the reasons in those other campaigns.
So, the "name-calling" is there - to help people remember the issue.
Is FUD present in any of these campaigns? "FUD" is a derogatory name for negative statements that drum up false fears, uncertainties, and doubts. In these various FSF campaigns, the fears, uncertainties and doubts are all quite realistic.
Dave Winer:
occupyweb.org
http://scripting.com/stories/2011/10/03/relyingOnFacebookTwi...
http://scripting.com/stories/2011/10/09/bootsOnTheGround.htm...
"small compared to SF, Palo Alto, or the South Bay"
That may be changing: http://berkeleystartupcluster.org/
Which reads, in part, "Launched in October 2010, the BSC's goal is to establish the area in downtown Berkeley -- that is walking distance to the Cal campus -- as a thriving destination for technology-oriented startups as well as tech-oriented established companies, investors, entrepreneurs and supporting businesses."
Since you know Berkeley you must know the "Power Bar Building," right? (One of the two high-rise office buildings in Berkeley.)
BSC has taken over the very top floor (gorgeous views) of the Power Bar building to create an incubator space.
My zoning idea would open up street level spaces for start-ups to grow into.
SF is very accessible from Berkeley, especially downtown Berkeley. BART is right there, for example, and by car the Bay Bridge is a quick hop down the road.
Palo Alto and the South Bay generally are less accessible. That's more of a trek either by public transportation or car. Here's the thing, though: across from the southern part of the peninsula (the traditional core of Silicon Valley) there is the south part of the east bay (e.g. Fremont) where high tech businesses locate. Between them is another traditional hub in San Jose. San Francisco has become more popular and has anchors like Twitter and a tight relation to Oracle. The bay is nearly encircled by hubs with the Berkeley / Oakland part being the laggards.
Berkeley does this because they want the retail corridors to be destination spaces with lots of foot traffic. Concentrating retail use like that creates a network effect: while you're downtown buying some stationary, why not get a coffee next door, pick up a book at one of the last standing independent booksellers, pop in the game store to buy a present for your nephew etc. If retail is shut out by office uses, that network effect is threatened and there is less reason to go buy stationary in the first place.
The landlords (some, at least) are saying that part of the problem is the physical structure of this retail space. A lot of the buildings are quite old. The retail spaces are "deep" -- 50' back from the front of the store, much of which was intended for use as "the back room" where inventory would be kept. Few retailers need that much space. Modern supply chains reduce needs for extensive inventory. If a retailer only needs 15' deep of space, it's hard to convince them to pay the same square footage rate for the entire 50' depth. Additionally, there is less demand for traditional retail at all because more products are bought on-line or at big-box stores elsewhere.
My proposal addresses this. It addresses the lack of demand for retail at all by encouraging cultural use like galleries and artisan shops. It addresses the "too deep" problem by allowing for offices in the back, opening up to cultural use in the front.
I think this could be a good environment for start-ups. If a number of inexpensively sublet cultural up-front spaces were active, it makes for a stimulating and attractive environment for start-up workers. Meanwhile, the semi-open offices-in-the-back would be more comfortable and interesting than a sterile office-park type office.
Wow.
The comparison to Albuquerque makes no sense to me. We could as well guess that Airbnb is less safe than riding a roller coaster at Disneyland -- so what? Those kinds of random comparisons prove nothing.
Because of its inherent structure, we can reasonably suspect that Airbnb is far, far riskier than current measurements yet show. I think of lots of nasty crimes that it facilitates and I'm sure you can as well. There is next to nothing in it that would prevent much higher rates of crime. It's structure is attractive to crime. The toll of the kinds of crime it is well suited for is incredibly high. Unless the basic model is significantly changed, none of those risks will go away or much diminish. Meanwhile the attractive potential for abuse of Airbnb is now well advertised to all potential criminals.
I suppose what most greatly offends and alarms me about Airbnb is how sharply it diverges from very well established cultural and legal norms. It is as if the designers of this product want glibly to overturn common sense about how far to trust strangers and it wants to ignore the many rules and regulations that are designed to prevent precisely this kind of business activity -- all on the basis that "most people are basically good" or something. That's socially irresponsible engineering.
What is their guarantee to guests? What prevents them from being a target for scams to collect host insurance? What statistics can they provide us about the wellbeing of all past guests and hosts?
If I'm going to rent out a room or a flat for the prices they advertise -- but apparently I better have an extra $50K of insurance before doing so -- is the two or three digit per night rent really worth it? This new guarantee sounds like an admission of failure.
I'm also quite sincere, wearing my "engineering hat", at least as I understand it, that there are serious life-critical mistakes getting papered over here.
a) Airbnb needs to recognize the risks they are creating for guests - this move is purely a reactionary response to placate public perception about the risks to hosts. They do not care about safety so much as the supply of the product they broker.
b) $50K is a pretty good prize for anyone who can game them on this one.
The "apology" says: "There have been a lot of questions swirling around, and I would like to apologize and set the record straight in my own words. In the last few days we have had a crash course in crisis management."
If crisis management is something you're only just now learning about, in this highly risk inducing business, what does that say about your founders, investors, and your investors' funders?
I honestly now wonder if there aren't already any guest victims who we haven't heard from because, well, they're gone. I seriously wonder if this company hasn't already killed someone.
More importantly, though, the scope of the risk of airbnb is enormous and past performance is not a reliable indicator of future success. As these two examples of bad experiences have shown, the system can be gamed. As a larger and larger number of bad guys learn that this system can be gamed (and how) -- exploits as a percentage of use and percentage of value received (insofar as you can quantify that) will, it seems to me, almost certainly grow.
The fraction of people with enough sophistication and resources to be good, safe, hosts and guests is a tiny, tiny fraction of the people in the world. It is absurd to think that airbnb could ever scale to justify its investments if only those sophisticated users were customers.
The model relies, therefore, on large numbers of hosts and guests taking foolish risks. Airbnb's light and airy "how it works" video is a fine illustration of just how glib they are in encouraging people to make bad mistakes. The advice on their "safety" page is risible.
Because most people are decent, most foolish risks from using airbnb go "unpunished" but certainly not all - as we've seen. A significant percentage of those risks lead to great loss.
I fear that Airbnb is going to help kill, kidnap, rape or otherwise grievously harm someone's person, directly, at this pace. That's because to make its numbers the company must, needs be, incite dangerously foolish behavior from many customers.
That the investors controlling so many dollars signed off on this investment speaks poorly of the VC system that produced airbnb. It speaks poorly of their common sense and/or dedication to basic social responsibility.