HNHacker News
TopNewBestAskShowJobs

jomohke

508 karma · joined March 4, 2009

submissionscomments
jomohke··on Coding Horror: Your Internet Driver's License
You can be fully in control of your own OpenID if you want. Just host it on your own domain/server.

Or even easier: use delegation. Add some meta tags to the header of your blog (or any HTML page under your control), and sign up for sites using your blog URL as your OpenId.

See an example at my domain (the content is currently blank, but the meta tags are in the source code): http://jmh.id.au

I can sign up anywhere by typing in 'jmh.id.au', and it uses myopenid.com for login. If for some reason I didn't trust myopenid.com I can change providers by editing that page. So I'm not tied to any one provider.

jomohke··on Open ID Is A Nightmare
Agreed. Though it would only work if email providers also became OpenID providers (which admittedly would have been very possible), or else having two things that look just like email addresses could cause similar confusion.

An advantage of URLs is that you have the choice of whether to give your email address to the site. Myopenid.com will prompt you for whatever information the site requests, and you can even choose which email address (if any) to provide the site. Sites which use an email address for login currently require you to remember which spam email address you gave them.

The indirection offered by URLs is quite powerful too: See jerf's comment: http://news.ycombinator.com/item?id=1916033

jomohke··on Open ID Is A Nightmare
OpenId does have this in the form of a URL as your identifier. If you had a myopenid.com account, for example, you can login to your site using 'michaelchisari.myopenid.com'. That page will contain a metadata tag specifying who your provider is (You could use any domain/page you control as long as you add a metadata tag to it).

This is the default way that OpenID works -- The selection of providers is purely a UI choice that has become popular by sites to hide users from URLs.

There's always still a box near the bottom that allows you to type in your URL without picking a provider.

jomohke··on Google offers JPEG alternative for faster Web
Have they explained their reasons for not backing JPEG-XR instead? This seems to be a step back from JPEG-XR:

- No alpha support

- Not lossless support (Useful for the alpha channel, and could encourage more cameras to support lossless images)

- No HDR support

JPEG-XR also allows different regions of the image to be encoded independently. See wikipedia for more features: http://en.wikipedia.org/wiki/JPEG_XR

I have no idea what the patent landscape for JPEG-XR is, but I'd be disappointed if we replaced JPEG and didn't get some of these features.

The lack of alpha support in JPEG is especially a pain for web developers. PNG does not do photos well.

jomohke··on Minecraft creator banks $350k a day, turns down job offers from Valve and Bungie
It's pretty difficult to do language performance comparisons based on a single game. Dwarf fortress, another great indie game involving mining and crafting, is written in C++, has no 3D (it's simply a 2D grid), and uses up a lot of CPU.

That doesn't mean that C++ is slow. It means we have no idea how much these custom indie game engines have been performance tuned, and how much logic is being performed in the background (DF is a hugely deep game). They are both written by a single programmer, not a team of game engine developers.

Edit: I don't doubt that performance tuned C++ is faster than Java. But for a small development team, who aren't building a graphically taxing game, performance optimisation is well down the list of priorities. They are usually better off spending the time expanding their game.

jomohke··on Welcome to the new decade
He's not talking about desktops: Apple still has tiny (but growing) market share there.

He's talking about mobile platforms. Microsoft is very much the underdog after ignoring the platform for so long. Apple had a monopoly over good smartphones by beating everyone to market by such a large margin, though it's hard to call them a monopoly now that Android has been selling so well.

jomohke··on Microsoft changed IronPython, IronRuby and DLR licenses to Apache 2
As far as I know, they haven't released C# 3.0 or 4.0 to the ECMA. I'd love to be proved wrong. See wikipedia: http://en.wikipedia.org/wiki/C_Sharp_(programming_language)
jomohke··on Ask HN: What is the best Java IDE?
I would phrase that question as "Am I editing one file, or a project of files?". As a Vim user, who previously only used IDEs for java projects, I now use IntelliJ for HTML/CSS/Javascript projects too. If you learn the refactoring and completion shortcuts, I believe it beats VIM at most multi-file projects.

If you install the Ruby plugin (also built by JetBrains, for their Ruby IDE) you get support for Sass - and they've just announced support for Less and Sass3 in the next version, including command completion and jump-to-declaration. http://blogs.jetbrains.com/ruby/2010/07/whats-mining-leaner-...

The Scala plugin doesn't seem too bad either, but I haven't used it extensively. It already has some simple refactoring features.

They're also working on a Python IDE, which is in beta, so the Python plugin will get fairly good too.

jomohke··on Ask HN: What is the best Java IDE?
We've used Netbeans and Eclipse for previous projects, but moved to IntelliJ last year.

The language integration is much richer - and it has a more solid foundation for working with multiple languages. For example, find usages and refactor-rename will work across different languages and templates. When I find usages for a field, for instance, it shows me where it is used in hibernate queries, EL expressions on pages, etc.

Languages can be arbitrarily nested in each other, a lot more so than other IDEs I've used - for instance, I can still use command completion and refactoring in nested SQL, HTML, CSS or even Javascript (eg. GWT's JSNI methods) when it is inside java strings, or XML documents, or wherever else.

The JavaScript support far exceeds either Eclipse or Netbeans. IntelliJ has fairly decent refactoring and completion for Javascript, which the other IDEs (at least as of last year) barely supported.

It has full completion, find usages, and jump-to-usage within JPA/Hibernate queries, which all ties straight back to the original entity classes.

The refactoring tools are a lot richer, and a lot faster than Eclipse or netbeans offered. On a largish project the refactoring commands still feel near-instant in IntelliJ -- it may not sound like such a big deal, but after becoming accustomed to it, I use the refactoring tools constantly while writing code. I didn't do it nearly as much with the other IDEs.

The refactoring commands are all extremely solid -- for instance, the extract method refactoring will detect similar code snippets and ask you if you want to replace them with the method call. I never had Eclipse or Netbeans ask me that (Have they changed in the newer versions?)

I've also found the support for Git far exceeds the other IDEs.

While the price seems expensive compared to the free competition, most programmers use their IDE 8 hours per day -- given programmer wages, it only takes a tiny improvement in productivity to pay for itself.

jomohke··on Microsoft by the Numbers
Yep, this always bugs me too. When it came out it had quite a good rendering engine compared to most (all?) competitors.

The problem arose from Microsoft disbanding the IE team and not releasing a new version for so many years. It fell behind from age.

jomohke··on Less.js will obsolete CSS
The latest version of Sass (3.0) has CSS syntax: The compiler supports both variants - the file extension (.scss vs .sass) denotes which you want to use.

See the examples on their front page: http://sass-lang.com

They have a command line tool "sass-convert", which can convert between the syntaxes (and from Less)

jomohke··on Less.js will obsolete CSS
We used Less for a project, but ended up moving to Sass: http://sass-lang.com/

The Sass compiler was (is?) more mature, and the language feels more consistent (see the '&' operator in Sass vs the :hover special case in Less). The editor/tool support was quite a bit better too.

Sass doesn't have a javascript in-browser implementation (afaik), but for rapid development it can watch a folder/file for changes and automatically recompile to static CSS files.

Plus there's a firebug plugin, so you get Sass line numbers / references instead of compiled CSS line numbers: https://addons.mozilla.org/en-US/firefox/addon/103988/

jomohke··on WWDC10 videos are here and are available free.
Any recommendations of good talks? A session number is enough info to avoid breaking the NDA.
jomohke··on Firefox 4: An early walk-through of IndexedDB
The other common complaint is that using plain strings for data access is inelegant, and not very javascript-y. It will face the same problem as other platforms, where developers end up writing ORM frameworks to map relational database tables (somewhat awkwardly) to objects.

IndexedDB feels closer to real javascript - it does all access through javascript methods. Writing storage wrapper libraries will be cleaner against this kind of API than concatenating strings of SQL.

And it does not preclude someone from writing an SQL implementation on top of it, if they desire. (Or something simpler. eg. http://thomasfrank.se/sqlike.html)

jomohke··on Lightspark: LLVM-based Flash Player for Linux
The two projects appear to have implemented different parts of flash. Flash has several separate virtual machines, which have been introduced over the years with new versions of flash.

Lightspark only implements the newer ActionScript3 runtime, from Flash 9 (and newer), while Gnash only (?) implements some of the earlier runtimes.

Currently Lightspark can only run the newer HD youtube player (which is compiled against flash 9), and Gnash can only run the older player (Which I think is still used for VP6 videos)

jomohke··on The ARM, the PPC, the x86, and the iPad
The article a little out of date. The RISC vs CISC distinction is very blurry these days. Most RISC architectures have been made more CISCy, and vice versa.

The article mentions CISC CPUs using RISC instructions internally, which isn't the whole story anymore.

Ars Technica did a classic article on this debate (in 1999!): http://arstechnica.com/cpu/4q99/risc-cisc/rvc-1.html and a follow up here: http://arstechnica.com/hardware/news/2009/09/retrospect-and-...

A particularly interesting part of the follow up: "But a funny thing happened with the Pentium M: processor designers discovered that processors of all types are actually more power-efficient if their internal instruction format is more complex, compound, and varied than it is with simple, atomic RISC operations ... ... The end result is that even RISC processors needed to get more CISC-y on the inside if they wanted to juggle the largest number of in-flight instructions using the least amount of power."

jomohke··on Bye Bye Flash: GitHub's Network Graph is now Canvas
I'm surprised he didn't use SVG - it provides an object graph, and will handle clipping/rendering for you.

A low level API like canvas probably isn't necessary for this kind of work.

jomohke··on Google Closure: How not to write Javascript
Agreed. I was expecting a critique of the library API. JavaScript APIs (including, especially, the browser APIs) have a long history of being overly-verbose, and designed as if it was Java.

Has anyone used the closure library API enough to comment?

jomohke··on Ask HN: If C++ is so bad, what should game developers use?
The D programming language [http://www.digitalmars.com/d/] was designed primarily to be a better C++. It has C-style syntax, a vaguely similar OO model to Java, and compiles directly to machine code, with performance as priority (it even allows you to drop to assembly if desired). The lack of C backwards compatibility allowed them to clean up a lot of the uglier areas of C++.
jomohke··on The C++ bashing season is back
Such a distinction between good and bad parts is much more difficult to define for C++ than Javascript. Javascript hasn't had the problem of programmers using separate subsets of the language. The bad parts of Javascript are design/implementation mistakes. The bad part of C++ is the overabundance of language features/complexity.

However, I think it would be great for someone to design a sensible subset of the language. Provide a lint tool (like Crockford's JSLint) which can enforce the subset. Marketing will be the hard part - a buzzword/name will be extremely useful for branding/word-of-mouth to work. A name will make it sellable, and identifiable, as a language of its own.

jomohke··on IntelliJ IDEA goes open source
Yeah, IDEA's javascript/html/css support is one of the reasons I switched from eclipse. The Javascript editor in particular is much improved over eclipse: Autocompletion, rename refactoring and find usages actually work somewhat!
jomohke··on Poll: What do you think of OpenID?
Use another OpenID provider, such as https://www.myopenid.com/

Or you can host your own.

See here for a good list: http://openid.net/get-an-openid/

I initially wasn't too impressed by OpenID, but I was sold by the third site I got to use my login. It's value is really tied to the adoption by sites.

It's especially nice on small blogs: you don't have post anonymously if you are averse to signups.

jomohke··on Gcc Misinterprets Trivial Loop as Infinite
Many complaints revolve around GCC's monolithic architecture. See the clang comparison page (http://clang.llvm.org/comparison.html) for a decent overview.

clang should open the door for intelligent refactoring and analysis tools which aren't possible using the GCC codebase.

jomohke··on MacBook Pro, Thousands of Colors
Apple is changing this in Snow Leopard: The default gamma will be the same as windows.
jomohke··on Working on It (TextMate)
I think his point is that it isn't Cocoa itself which makes the new version faster and more extensible: It's the rewriting process. The new codebase has presumably been better structured to be faster and more extensible.

As John Gruber has said "Cocoa is just an API. It is not some sort of magic technology where you just sprinkle a ton of square brackets in your source code and you instantly get a better UI." (http://daringfireball.net/linked/2008/10/30/cocoa-finder)

← PreviousPage 4 of 4