XMMS2

New XMMS2-Scrobbler release. Better than ever: multi-scrobbling is supported now. libre.fm users rejoice! See README for the details.

For a long time I haven't been able to develop anything constructive. It's been a severe code block that has been lasting for years. That's why it so fun to now announce some of my recent projects!I decided to really learn Objective-C back in ... [More] November. At work a customer requested a iPhone application and I was the only one that knew anything about Objective-C, so I was asked to join in a customer meeting telling them that I really knew Objective-C and iPhone development. So instead of lying my ass off I sat down and tried to get some coding done.Eventually the customer didn't want to pay for the application so the project was never started, but I really learned Objective-C. And it resulted in my first iPhone application! It's a dedicated client to the MobilBlogg community. Together with Henrik Öhman of MobilBlogg (who did the API) an application to browse, upload and comment on photos was created. The application is submitted to Apple, who knows when it's going to be accepted :-)The application is Open Sourced under the GPL license and you can check out the sources here.The application could never had been made without other brilliant open source code: Three20 and JSon Framework. If you want example of using these frameworks, checkout my code.To view screenshots and photos uploaded with the application head over to my mobilblogg! [Less]

For some mysterious reason, Internet Explorer (tested in IE7 and IE8) strips leading spaces in text nodes preceded by an empty element, such as this: [code lang="html"] foo [/code] While innoccuous in static pages, it becomes problematic when DOM ... [More] nodes get updated after a delay by some Javascript, as the separating whitespace has disappeared, hence ruining the layout of your text. Surprisingly, I haven’t found any reference to this IE bug (not that there is a shortage of complaints about other IE idiosyncrasies), so I thought I’d share the problem and the solution I have found here. It’s all demonstrated in the following example: [code lang="html"] !DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd"> world   world [/code] In the first case, the whitespace gets stripped, thus resulting in “Helloworld”. In the second example, the non-breaking space prevents IE from stripping the whitespace, so that when the span is filled by Javascript, the text reads “Hello world” as expected. If you know other solutions or whether it’s considered a bug that will be fixed, please post in the comments! [Less]

On my way to the Google Summer of Code Mentor Summit 2009, I accepted DraX’s invitation to give a 1-hour talk about XMMS2 Collections at his work, i.e. Metaweb, in San Francisco. The topic was somewhat relevant for them as it’s reminiscent of MQL ... [More] , the query language they developed for Freebase. It’s worth noting that although both share a pool of buzzwords such as “graph”, “loosely-structured”, “querying”, etc, they are not quite the same: Freebase is essentially a giant graph-database, which you query with MQL to retrieve graph fragments. The XMMS2 database is a flat denormalized store, which you query with graph-structured Collections to retrieve a list of entries. Note: Collections 2.0 should however allow fancier querying to retrieve tree-shaped structures. About 10-20 people showed up and listened to me blurbing about the concept of Collections, the rationale behind them, the API, Collections 2.0, possible UI uses, what it represents for the user, pointers to S4, etc. It’s all in those over-engineered slides that I have no choice but to put online, under Creative Commons Attribution-Share Alike 2.5 License, for them to live on forever on the internets. And yes, it’s still either in evil Keynote format (source), or in PDF. Oh and Metaweb, thanks for the food! [Less]

Thanks to Google’s everlasting generosity, mentors from all Open Source projects that participated to the Google Summer of Code 2009 have flocked en masse to Mountain View to meet up and talk about code and drink beer. More in pictures in my Bay Area ... [More] Flickr set, if you’re curious. Sessions were organised spontaneously around various topics, and one of which, proposed by Amarok2’s Lydia/nightrose, was “Problems Audio Players face today”. Which resulted in 15-20 people from various projects (incl. Amarok2, Rockbox, Maemo, XMMS2 & others) talking about solving problems encountered by all music player developers (e.g. lyrics and cover art fetching), as well as features on our beloved users’ wishlists (portable player support, tags, musicbrainz, etc). And if you weren’t there, don’t despair, for I just (finally) posted minutes for this session on the GSoC wiki. Some interesting ideas in there, so go and have a look, and get to work! [Less]

We are planning to do a XMMS2 Con next year. Please visit this page to announce your interest: http://wiki.xmms2.xmms.se/wiki/XMMS2Con2010

You've probably already heard of Sartak's awesome MooseX::Role::Parameterized. As of version 0.25, MooseX::Declare provides some nice sugar for that. This is what it looks like: use MooseX::Declare; role Counter (Str :$name) { has $name => (is ... [More] => 'rw', isa => 'Int', default => 0); method "increment_${name}" { $self->$name( $self->$name + 1 ); } method "reset_${name}" { $self->$name(0); } } class MyGame::Weapon { with Counter => { name => 'enchantment' }; } class MyGame::Wand { with Counter => { name => 'zapped' }; } Just thought I'd let you know, since i haven't gotten around to actually document this. In case you like it, have some spare tuits, and want to help out, the repository is right here :-) [Less]

IEEE Software recently published an opinion piece from Tom DeMarco titled ‘Software Engineering: An Idea Whose Time Has Come and Gone?’ in which DeMarco appears to declare the death of ‘Software Engineering’ as a discipline. Unsurprisingly, this is ... [More] generating some waves in the tech blogging community, with pieces such as ‘Software Engineering: Dead?’ from Jeff Atwood’s Coding Horror, and Myles Eftos’s ‘Yep, Software Engineering is dead’. From reading DeMarco’s article, it is apparent that what he is addressing is the emphasis placed on ‘control’ on software projects - that ‘software engineering’ is often taken to mean that it’s a means of controlling the outcomes of developing a software project. It should be obvious that DeMarco’s stance is one of reflecting upon his previous work and making some comments about how his work has been perceived in the software industry over the years. Aside from some “Aw shucks, I wish my younger self had been wiser” style reflection, DeMarco laments the fact that his work has been used over the years to justify the creeping bureaucratisation of the software development process. From that point, DeMarco goes on to prescribe that, rather than controlling minutiae, those in charge of software development should be treating their projects (by way of analogy) as unruly teenagers - by attempting to instil a sense of moral values and principles, rather than by being overbearing and all-controlling. Unfortunately, Jeff & Myles appear to either be sensationalising, or reacting to a rather superficial reading of DeMarco’s opinion. Straight off the bat, Jeff points out DeMarco’s gravitas in his field, to hammer home the point that ‘yes, this is a big deal’. From there, however, Jeff switches to his usual agenda of elevating the romantic notion of ‘software craftsmanship’. Not that I’m averse to this position - au contraire, I’m a big fan of Jeff’s frequent expounding on matters of software development principles. What I disagree with, however, is the utter sensationalism with which this is being made here. At least, I am heartened that Jeff appears to have grasped the main point of DeMarco’s analogy: “If you want to move your project forward, the only reliable way to do that is to cultivate a deep sense of software craftsmanship and professionalism around it.” I have no idea what has been said about either DeMarco’s piece or Jeff’s blog post on the 220 mailing list, though reading through Myles’ blog post left me slightly incensed at certain points. The analogy between building a bridge and building software is somewhat tired. Compared to building bridges, building ‘software’ encompasses a vastly larger scope. That much we agree on. You might as well be comparing building software with building anything made of materials sourced from the ground. From here, however, it seems that Myles launches into a tirade about his least favourite academic subjects, or at least the ones that he found ‘stupid’. What I fail to grasp at this point is the relevance to DeMarco’s piece… But I’ll play along anyway. Yes, formal proofs are arduous, and can be rather silly for trivial programming tasks, but it should be obvious that such techniques aren’t applicable to all software development contexts. Would you apply formal proof methods to a small dynamic website you’ve just built? Hell no. Would you apply formal proof methods to mission-critical software that could potentially lead to injuries, fatalities or serious damage to the environment or property, if anything were to go wrong? Hell yes. (Well, some variant of a formal method, anyway) It’s a simple cost-benefit (or risk analysis) scenario: which costs less - the development process, or the consequences of something going seriously wrong? I’d further argue that formal proof exercises done in an undergraduate academic setting would be more geared towards teaching and learning the process of doing the proofs than anything else. This is undergrad CS that we’re talking about, with its endless assortment of silly assignments that require students to produce little of value beyond the end of their courses. I might also add that, even with a ‘simple’ program having 9 lines of code, this degree of simplicity may be afforded by a myriad of much more complex underlying supporting systems - the degree of simplicity of a program may simply be a matter of perspective, or from where you’re standing to look at the code. What I find most troubling about Myles’ post, however, is the complete denigration of UML. Yes, I’ll admit that I’m a big proponent of UML. There may be a point about UML working really well with the waterfall model of software development, but to then imply that UML is almost completely useless - by association with the inadequacy of the waterfall model to address ‘real’ problems - seems a lot like missing the point of what UML is actually about. Yes, UML may superficially look like ‘architectural drawings or flow diagrams or something’. More importantly, UML is a standard notation to communicate about elements of software design. Beyond the ability of expensive software to create specky, neat and professional looking UML diagrams, UML itself is a tool for software developers to use when thinking about, crafting and communicating about the systems they’re building. By analogy, I would say that an ability to work with UML is like an ability to work with words and sentences and paragraphs. That is probably the reason why IT recruitment agents started using ‘Object Oriented Analysis & Design Skills’ and ‘Sound understanding of UML’ in their job ads - those *are* fundamentally useful skills to have. As a programmer, I can have a conception of something that is a ‘class’ or an ‘object’ and be able to express the idea in a diagram, say on a whiteboard, while talking with fellow programmers, and still be able to more or less directly translate this concept into code using Java or Python. Using UML, I can model relationships between such classes and talk about such relationships - beyond mere representation, UML can be a valuable tool for solving software design problems, especially in team environments. [And the good news is that there are tools nowadays to automate parts of the translation between UML and code, so you don’t have to ‘go and update hundreds of UML documents every time a minor change was required’] None of this is really dependent upon the specific project management methodology being used. The real problem, as far as IT recruitment is concerned - at least in this region anyway - is that there is often a disconnect between what software development shops need and what their recruitment agents end up publishing. Given that IT recruitment agents are often more experienced in HR and marketing, they end up developing formulas for easily pushing job ads out there seemingly without much consideration for what is really required. That is how you end up with a litany of job ads that look like the same stuff rehashed over and over again. Thanks to all this, graduates who do have such skills end up in jobs which are advertised as requiring their skillset, yet never use them in any useful capacity. Getting back to the main article, however - the point that DeMarco is trying to make here is rather subtle, one which can best be understood in the very context of his analogy. Developing software, in general, applies to a vast field of human knowledge and applications. As such, decisions about how software should be developed necessarily need to vary based upon the context and constraints within which such development is occurring. By way of analogy, humans, as they grow into adults, cannot be robotically programmed to gracefully handle every possible situation they might encounter. (Or, at least, one could imagine that this approach could fail rather disastrously if tried) Given these limitations, it is far better to learn about core principles and knowing how and when to apply them to produce generally favourable results. DeMarco’s point is that, as this applies to humans, that may be the real way forward for software development. p.s. Some interesting and relevant articles relating to this topic: - “Is Software Engineering Engineering?” - Peter J. Denning, Richard D. Riehle - “Software Engineering ≠ Computer Science” - Chuck Connell [Less]

Over at The Digitalist blog there is a very good discussion about eBook DRM. Cory is excellent as usual but the highlight must be Clay Shirky's quote:You say “I’m a strong believer in the rights of creative artists to control the distribution of ... [More] their copyrighted content.” Sure, why not? And while we’re at it, we can be strong believers in breathing under water, or living to 130. Those ideas are similar to the idea of controlling distribution, in that there are a whole bunch of people who think they would be really really great. They’re all similar in another way too. Can you guess what it is? [Less]

What’s calypso? The new official xmms2 gui client. I could talk about why the name calypso is a good name, what it means, etc. But I won’t, I’ll write about technical and implementation details. Based on theefer’s post about extensability and some ... [More] talks we have, we get to the conclusion that the client should be extensible, based on this I implemented the prototype of a framework to be used during this GSoC and hope that it will be useful in the final client. The framework is composed of modules, there are two types of modules, I will call them the core and the extension modules. As you have thought the core modules contain core features and serve as the base for the construction of other modules. The core modules are responsible of communication with xmms2 daemon: to control playback, retrieve information, change configuration values, etc. They’re responsible for maintain caches of retrieved data and provide interface to this functionality to other modules. Actually if you look in the code tree you will see three (partially implemented) core modules: playlist, playback and collection. Let’s talk about the more interesting, extension modules. In this category falls all modules that does not interact directly with the xmms2d daemon or provide basic functionality to the other modules, normally modules in this category have some related UI, for instance, a playlist viewer, collection browser, playback controller, and what else you can think. I’ll post about the ones I implement. Another difference is related to implementation, core modules are supposed to be implement in some predefined default client language (C++), while extension modules could be implemented not only in the default language but also in some high level extension language (we’ve been trying QtScript). The current state is the following: there are core modules for controlling playback, playlist and colllection management. In the extension modules field there are a simple playback controller, a playlist viewer, collection browser and the calypso-bar (this one deserves a specific post, but basically, it’s a cli-like controller for all modules). In terms of Qt naming (I haven’t mentioned but the current prototypes are being developed int python+qt) models classes are normally provided by core modules while the extension modules are viewers and controllers (obviously it’s not a rule, for instance, there can be a lyrics fetcher extension module that’s a model). Related to the UI a module is responsible for providing his own interface (if have one) and there are no restrictions in how it’s created (manually in code, loading an .ui file) it just have to follow some naming rules so it can be used in the main interface. I think that is it what I wanted to talk in this post, the next ones will be about some specific modules, in the queue there are already the calypso-bar and collection bins. The tree is in git://git.xmms.se/xmms2/calypso.git you can try it, but be warned for now it’s just prototyping, have lots of bugs and partially implemented things. If you try anyway, please comment. [Less]