aggronn

Thats actually what I had settled on using, and I think its a fine tool. But even umlet doesn't make writing and viewing large UML diagrams...pleasurable? In the sense that we have tools like vim, word, excel, and chrome that are very polished and allow you to get to work quickly and accurately. Even with umlet, i find myself worrying about the interface too much.

I suspect its more of a problem with the fundamental design of UML though. I think if there was like, a UML-lite spec or something. To me, i feel like UML is to what I want to use as XML is to JSON (or csv for that matter), if that makes any sense.

In school, UML seemed like a great idea. It wasn't very fun, but it was useful, and if you had a simple class project, UMLs certainly made things simpler if you were even okay at them.

When I got my first job, however, there were no UML diagrams. I was saddened to see this. I don't remember if I was surprised, but I was so sure that I would been useful to have a detailed UMI diagram available for studying the codebase.

So, as I got acclimated to the codebase, I took it upon myself to start putting together a UML for our existing product, so that training new people would be easier. After spending several days doing this, my opinion or UML changed dramatically for a couple of reasons:

1) Pretty much all of the UML tools out there are not a pleasure to use. Drawing UML diagrams was still infinitely better than using the various programs I tried. But drawing them on anything but a whiteboard wasn't scalable to the size of the project.

2) Once I had the basic classes drawn out, and realized how diverse the kinds of classes we had were, it became apparent that this just wasn't going to happen. A basic flowchart of the application would be equally valuable. The specificity of UML wasn't worthwhile, having a class with 10+ subclasses was a mess.

3) What I did have was hard to see. This goes back to the first point--the tools are not good. When you labored over all the clicking and dragging, you pulled together this monster of a graph that you can only see a small portion of at a time in a given screen. It was easy to get lost. Printing wasn't useful.

Ultimately, I only ended up using UMLs when we were thinking about how we were going to build out new features, or redesign stuff. It was a .NET shop building a Silverlight app, so serialization and strict static typing was critical. A decent JSON api would have made life easy, but since the back end and the front end were essentially all the same thing, class diagrams were very useful. So for the 2 or 3 times that we had to develop or redo class serialization while I was there, it came in handy only for working on small 2-5 class models.

That brings us to the next point: do class diagrams have a place in an increasingly dynamically typed world? Functional programming in javascript. Frameworks that have their own detailed class system, or which aren't even class oriented.

UML has essentially become analogous to cursive. You learn about it early on, and its important to have a notion of how to read it. But ultimately, since we don't use it every day, we basically never use it.

Honestly, if total surveillance is constitutional, you have to figure the constitution isn't really that good.

it is ruby-doc. What don't you like about it specifically? Its very concise, consistent, and thorough, which is what I always thought was ideal for these.

This has to do with a manicpal law, so no. If another city had an identical law then maybe, but airbnb is plainly in violation of it--the court just triple checked.

I like how he mentions how good job prospects are, even for mid-level schools, then cites Michigan. This article is worthless.

edit: to clarify, as mentioned below, Michigan is a top program.

You're right, I ignored the depreciation on the remainder of the mortgage and assumed that when you sold the house, there would be no loss on that.

Incredibly stupid example. Embarrassing really. Objection withdrawn

I missed the latter part of the title. Can we have NSFW stuff labeled at the beginning of the title since they're so uncommon and unexpected?

Grooveshark doesn't do any audio ads and never cuts you off.

Honestly, I think this is a brilliant idea for a few reasons, but the most significant can probably be explained by this real-life comparison:

Consider two scenarios. a) You go out to a party, by yourself, with the sole intention of meeting someone you can take home, or ask on a date. or b) your friends think you need a girlfriend or boyfriend, and drag you out to a party and try to get you to meet people and facilitate it that.

Scenario (a) is a relatively uncommon occurrence with, what I assume to be, a lower success rate than (b). Unfortunately, all successful dating sites thus far are almost perfectly analogous to (a).

The differences between the two situations are pretty big because of the countless nuances that make people just generally seem more sociable when they're being social compared to when they're sort of trying to be social.

In a sense, this introduces the concept of a good wingman into the online dating scene.

Beyond that, I think by having other people manage your profile, you're taking away some of the stigma associated with dating profiles. Among them that they can become very self-important, uncomfortably personal, and come with a certain level of expectation of being honest. By making everything second-hand, you have to be more realistic in your evaluations of people's profiles.

Moreover, I think it dampens the 'desperation' factor. I think this is also a lesson we can gain from the analogy in my second paragraph.

Why? He could have sidestepped any plea bargaining if he wished. I'm not sure what you're suggesting they do.

This has been a very common theme through my undergraduate stats education.

Okay, thats one app. What about all the other innocuous apps you download?

I'm not following this thread; Are you suggesting that him sharing tabs isn't subject to copyright? Tabs sharing websites have to pay royalties, and have had to for several years:

http://en.wikipedia.org/wiki/Tablature#Legal_issues

Have you tried guitar pro? It doesn't have any video component, but as far as tab software goes, this isn't even comparable.

but as far as html5 tools for guitar tabs goes, this is certainly a huge step forward.

Consumers usually have an expectation of quick turnaround, and when they don't have a quick turnaround, they expect good feedback about when to expect what they paid for to be delivered to them.

In light of that, no, we don't really have the technology to automate all of those business decisions (especially so early in the product lifespan) and give a precise prediction about when a consumer can expect the product with acceptable reliability. There are too many factors that can delay delivery by a day or two, and accuracy within a day is extremely important for good customer service.

For a mass-market phone like this "It'll get there when it gets there!' is not an acceptable response to people who need a phone right now, or in the very near future. This isn't a raspberry pi.

There's no use pretending that high tech supply chain management is a simple process, and any hickups are really just ploys for media hype.

'Theoretically', no. Its hard to be optimistic about these two ideas because while the chess example is a good story, its not analogous for many reasons, ranging from disparity in available information to players to a difference of several magnitudes in saturation. Not to mention HFT just isn't chess.

HF traders are just as much hackers as anyone on HN (and there are plenty of HF traders on HN). So 'theoretically', they've already done what is being suggested here. If someone comes along and develops a winning strategy, it really shouldn't be considered as having anything to do with 'professional strategy vs novice strategies'. It would just be about one person either getting really lucky or coming up with something that is genius in its own right.

--

If there are 'professionals' and then other 'professionals' whose strategy depends on information about how other 'professionals' trade (and there is), you end up with strategies at all valid points in the sample space of possible strategies and counter strategies. Theoretically, there should be no other possible strategies. Inevitably someone will come up with one though, and the 'sample space' will grow. But its extremely unlikely that additional unique strategies are successful just because they 'counter' the strategies in the sample space. But then again, this is real life and these things aren't impossible.

Very impressive. Definitely the best data presentation dashboard I've seen. The depth of information (even if the source of information is so centralized) plus the diversity and clarity in presentation, not to mention the overall cleanness of the whole thing, really is very cool.

Who said something about extended durations? These would be very short durations.

I don't think that's necessarily true, though I disagree with the GP. He mentioned in the video, and they're already demoing this at places like best buy, ultrabooks will be coming out with touch screens en masse once windowes 8 is released. A lot of them pivot, but after using one briefly in a normal mode, I found that I was clicking smaller things with the trackpad, but that for big things (like the start screen), it was much more convenient to quickly flick the screen than to go and use my dexterity on the trackpad. you wouldn't use touch in office, for example, but you might use it for navigating around the operating system quickly and easily.

Having the mindset that the mere existence of bugs constitutes a personal or professional failure is probably very unhealthy. Is this something developers experience? I've never felt bad about a bug before. I just go and fix it. No pain, just gain.

Breaking things and inconveniencing others are a much bigger drag for me. Those are the problems I stay late at work to solve.

Maybe I'm misunderstanding this product. Please let me know if I am.

$10/month for 5 users on IRC?

This service is very poorly branded. IRC is a protocol that is easy to set up, free, and low maintenance. Whatever value there is to outsourcing IRC server setup, it isn't $10/mo. But that isn't even what their main service is. Their platform is just built on IRC--their other features (Archiving, workflow integration, etc) are much much much more value-adding. Its like advertising heroku as "rails hosting, and more!". It doesn't point to the real value of the service.

I can get easy IRC hosting for much cheaper than $10/month. Don't charge someone $10/month for an IRC server. Charge them $10/month for a developer-oriented integrated messaging platform or something.

It was unclear if what you were implying was that you shouldn't violate a ToS for your own profit (or at someone's expense) because it was a federal offense. (I can see now that that was not what you were trying to get at)

What are your thoughts on PadMapper vs CL? What is the distinction between scraping that data vs scraping this data that makes one worthy of federal prosecution, but not the other? Considering in both cases it was done for profit or detriment

If a restaurant had a 'terms of service' that said no reselling of food that they make, should you be open to federal prosecution if you stopped by to pick up food for yourself and some co-workers? Especially if someone gave you a few dollars for the trouble of running the errand? And even more, if the restaurant has their own delivery service?

Am I missing an important distinction here? Should private companies be able to make binding rules that open people up to criminal prosecution for something that doesn't violate any laws per se? A person breaking a specific law AND breaking a ToS makes sense. A person breaking a law BY breaking a ToS doesn't make sense.

What?

As a former economics major, it pains me to have my lost love's name used to describe mundane finance calculations. My liberal arts education be damned; I know nothing of this madness!

This makes me wonder: which is more valuable? writing blog posts or reading actual news?

I agree--I was thinking of solutions to this, and the most immediate thing that comes to mind is that you have to be extremely proactive in capturing small groups of people who are already organized around each other. If you have a start-up community in a college area, for instance, go out of your way to facilitate student organizations (and don't just send out a few emails to potential users). If your community has the ability to organize events, and and the salsa dancing club secretary doesn't want to have to deal with setting stuff up, sponsor their club with money or do it yourself.

Does this work on the iphone? I expect this would be a great alternative to socket.io for mobile that doesn't support websockets or flash.

Without reading the context, I expect that 'computational modeling' refers to a largely mathematical model that is potentially agnostic about underlying causes. You can run simulations that accurately reflect the real world using one set of variables, then run another model (or somehow test another model) that confirms the original simulation's results from a different perspective; ie, you have two theories, and you validate the second by proving its equivalent to the first, even though it uses a different set of variables, or looks at it using a different theoretical framework.

of course, you're right that this can be troublesome. just because you can validate a theory as accurately modeling reality doesn't mean that the components of the model are true or useful.

and necessarily, the first 'computational model' would have needed to be already validated before something else might be validated against it...in case that wasn't clear.