I'm going on my eighth year or so of using it.  I don't know the exact timeline, but I'm pretty sure I've used it since around April of 2002 or so.  I'm also coming up on my fourth year of being a Gentoo developer.  That just blows me away.  I've also started recently getting a resurgence of interest in working on the development side again, something I'm really glad to see, since I have been progressively phasing myself out over the past few months.  Quite recently, though, I've managed to get myself much more organized and that has given me a much bigger picture of areas where I want to focus (and ones I want to drop interest), and so that has made it working on Gentoo fun again.   For the record, that's how it should always be.

There is one argument in particular I wanted to rehash here, and that is the one about stability as a distro.

I've often heard the argument, when debating Linux distributions, someone say something like, "Why would anyone run Gentoo .. it's not stable!"  I never have a really good retort to that answer when people bring it up, other than something simple like "Well, the obvious reason would be that I've used it for years and if I manage it, you won't have any problems."  Not to say that I don't have problems with my Gentoo installs, but these days they are usually stupid things ... like forgetting my root password or accidentally firewalling myself out of the box.

I don't really care about the "distro wars" though.  My motto has long been, "to each his own."  Find what works for *you* and apply it.  I could argue all day to someone about the merits of Gentoo, but really, all that matters is that it fits my needs well and does exactly what I want it to do, and how I want it to do things.  Everyone else can go find their own.  I've got mine.

But the fallacy in that original argument "Gentoo's not stable!" is a weak one.  It implies that the distribution should be the one providing all the support for keeping things constantly maintained, running and efficient.  To me, those tasks sounds like the description of a systems administrator job position.

My point is that, it shouldn't matter what Linux distribution you are going to use -- you need someone to keep it up and running.  I think Gentoo is great because it removes the veil from saying, "just run these versions of the software and you'll be totally fine."  Bugs creep in all the time.  Binary distributions stick you with a set of packages, that if,  you want to break out of that pigeon hole, it may be completely impossible to do.  With Gentoo, the definition of "stable" is left up to the user, the maintainer, the systems administrator.  I love it.

Go Gentoo.

I also can't believe I'm still a fanboi, after all these years.

So, the backstory is that the other day I was making a note of all the projects I have to do,  creating what I call my project tree.  It's basically an ordered list of general projects (Gentoo, home theater system, etc.), and then abstract projects underneath that.  It's nice because I can get a birds eye view of all the stuff I'm working on without going into specifics about any of them.  One of the home theater ones was to get at least ISO disc rips of the Blu-Ray movies I already have, even if I can't decrypt them yet, so I can at least play around with my options as they become available.

Well, I had moved my server setup around not too long ago, so I had two 750 GB harddrives just collecting dust.  I put one of them in an external USB drive, and plugged it into my PS3, which was already running Linux (see this pervious post for all the fun details).

I formatted my external USB drive as NTFS, so that I could read/write to it with Windows as well, and then I would insert a disc and just dump it to an ISO file.  That's easy enough:

$ cat /dev/sr0 > KFP.iso

I grabbed a couple of them (which took awhile, don't let me kid you on that part) so that I could get a good sampling in case I had more luck with one than another.

Once that was done, I trotted the little drive and plugged it into my netbook, running Windows XP, and installed both Virtual CloneDrive and AnyDVDHD.  The first one lets me mount an ISO as an actual disc drive, and the second actually decrypts the disc for me and dumps the contents back to the harddrive.  So, that's two passes now on all the data, which is making this take a long time.  But that's okay, it's fun.

Now that I have the actual contents, the MPEG2 transport streams, I can play it back with MPlayer, ffplay or whatever.  I'm having limited success with latest MPlayer on the files with VDPAU playback support, but it could just be my video card (GeForce 8600 GTS).  On my HTPC frontend, it works almost perfectly on there, with a nicer video card (GeForce 9300).  My desktop just craps out, though.  I can still watch it with Xv video out, though.

mkvtoolnix doesn't support m2ts files right now, so I don't have many options if I wanna change things around.  I'm still in a proof-of-concept stage, so I don't really care all that much.  Plus, my options are already limited.  AnyDVDHD is shareware that will expire in 21 days, and while it's amazing and works great, it's really expensive -- something like $200 for a lifetime license.  Eek.  With that, I'll keep trying my options on Linux.

The second piece of shareware I ran into (which also has a limited evaluation license, though this time for 30 days) is MakeMKV.  The Linux port is always a little more difficult to find, so here's a direct link to the details on the lastest version.  Boy, I never thought I'd be talking about shareware on my blog.

MakeMKV works really nice, too.  It's supposed to be able to do the same job as AnyDVDHD, I believe, but since I don't have the actual Blu-Ray drive, everything I've tried has never worked when just mounting the ISO direcly and trying to access that.  I believe that part of cracking the key involves having access to the drive.  I'm really not sure.  I've read a bit about the whole process, but it's still really confusing to me still.

Anyway, the software will let you access it directly through the decrypted contents, and that's what I did.  The interface is actually really simple and nice, and I would actually consider buying this one (it's much cheaper, at $50 for a lifetime license).  I'm trying to remember the last time I paid for a software license.  One that comes to mind, is that I actually have a valid registration key for Legend of the Red Dragon, the *really* old BBS door game.  Wow.  I think it cost me something like $15.

MakeMKV is pretty nice, though.  It snags the subtitles I select, and already includes the chapters as well as the HD audio formats.

So, that's it.  The whole process is pretty tedious, but it works, and I'm happy.  I don't really care about decrypting it *too* much right now, since I don't wanna go through the pain of trying to play them back over my HTPC just yet.  I'd need to do a lot of tweaks and upgrades to my system, and I really don't care that much.  It's not worth the hassle.  Especially, uh, since I just bought a new Blu-Ray player last month.

Really, though, I'm just doing this for the exercise since when I get bored, often times I'll want to play around with media files and formats and see what I can do with them.

Eventually I'll buy a BD-ROM drive and see what I can do, but for now I'm trying to save some $$$ and the whole point of this was to see if I could rip some discs with just the hardware available, and I could.   Woots.

On a sidenote, I posted aacskeys to the portage tree today, which is one of the tools users need to decrypt the keys on their discs.  Hopefully we can get some more hackers interested in poking at it.  That's always good.

Last but not least, here's an actual screenshot from the final rip.

I love TMNT.   Boy that's a post for another day ... which reminds me, I should get a copy of my home-made videos some day, that I made with my action figures.  Oh man, that'd be awesome.

First of all, I get a lot of ideas to do a lot of projects.  There's lots of cool stuff I want to do, and I have a hard time saying to myself "I have enough projects already in the works to finish, better not start another one," but I do anyway.  I tend to quickly overload myself sometimes that way, which can be bad for everything.  However, one thing I'm getting more strict on is only picking up projects that I'm sure I want to complete, that I'll see through until the end.  I very rarely, if ever, completely drop a project that I've started.  I will tend to put them on hold for a while -- sometimes years -- but I'll eventually revisit the idea (heck, the packages website is a perfect example of that).

I have a ton of projects I'm "working on," though.  So many, that I'm honestly afraid to write them all down for fear of being totally overwhelmed by the responsibility I put on myself for them.  I do, however, plan on getting them all done, and they circle around in my head on a regular basis, and often times I think of ways to integrate two projects (for example, adding an option to search gentoo planet(s) from the packages site).  I get a lot of interesting ideas all the time, but I really have to be careful not to overextend myself.

One thing I've been trying to do recently (as in the past year) is slowly shutter off some of the support I've been providing for the Gentoo tree directly, and ebuilds / herds I've in the past taken close care of.  It occurred to me way back when that it'd be a more efficient use of my time if I built out some project websites (like the packages one) rather than trawling the tree looking for ebuilds to fix, bump and repair (for example).  Not that I mind doing that, mind you, in fact I find it rather relaxing at times, but what's happened is that I've overextended my responsibilities again, and I'm trying to cut back.  Basically, my thought is that while I want to still work on Gentoo for a while, I don't want to make a career out of it.

Oddly enough, though, part of the reason I'm doing these community projects is so that I can more efficiently do other ones.  For example, at times I like to go through the multimedia packages and just check them to make sure we aren't missing version bumps, and go fix small bugs that I can take care of and just little stuff that isn't really important (in a sense of package popularity) but still relevant to a few users.  Those are fun.  But it'd make my life easier if I could more quickly track what has been neglected, more easily see what available version bumps are available (I still wanna hook into GnomeFiles and track their changes, for example), and stuff like that.  A lot of the tree-fixing stuff in Gentoo development is just monotonous, which is why it's hard to find volunteers to do it.  There's a good chunk of it that is just boring work!  And I'd like to help streamline that a bit.  That's one of my big goals.

With that goal in mind, a huge reason for doing the packages site was just so I can have a simple interface to get all the information I need, and finally a standardized set of data for categories, packages and versions.  That's mostly done, or at least the framework is, so now I can get going on the *really* cool stuff.  What I've done so far is really just the tip of the iceberg.

Anyway, I didn't wanna talk about just the packages site.  There's lots of other stuff I have going on.  It's interesting, even to me, to see which ones I'll want to juggle at a time.  I switch between them on a regular basis.  Sometimes I'll be working on the packages site, then my DVD ripper, then my scriptures stuff, then I'll work on theology ebuilds, then sound ones, then I'll look after ALSA, then mplayer, then I'll go back to tweaking MythVideo a bit, and round and round and round it goes.  I'm always working on *some* project, that's for sure.  It might do me some good to try and get a bit more organized, but I don't even do a good job of keeping track of bugs in my own projects.  I just track them internally for the most part.

So, I apologize for the epic behind status that I'm always in.  I'm starting to recognize more and more how much I'm holding people up on some projects, so I'm doing my best to gracefully exit those areas so someone else can come in and take over.  I'm still fumbling a bit at the best way to do that, but at this point in my life I have at least recognized the few areas that I'm sure I'm not passionate about anymore, and shouldn't be lazing around just pretending to commit once in a while -- of which, there are actually really few.  In fact, I can only think of one off the top of my head.

One thing that might be cool that I just thought of -- have a status indicator on my blog or something that displays the current project I'm working on.  That'd be fun.   Sounds like work, though.  I'm gonna go watch a movie.

The big thing I've been working on with the packages site now is making it more efficient.  The first step in that has been gathering some data on how often certain things are being called to see where optimizations are most needed.  So, the other day, I added a counter to the constructor of each class that would just tally each time the class was instantiated, and then I'd dump out the counter at the end of an import run.

One thing that surprised me is how often one particular class was being called -- PortageTree.  It's a really simple class, and all it does is set down some really simple variables that aren't going to change at all once they are declared, such as the location of the portage tree and it's metadata cache on the filesystem.  Pretty much used across the board on a lot of other classes that need to know the filesystem location of files (PortageCategory, PortagePackage, etc.).

Well, being still pretty new and fuzzy to the OOP approach, I thought it made sense to just extend the PortageTree class on PortageCategory and call the parent constructor to get the variables set.  That ended up in that class being created a huge magnitude of times,  all for the same pretty much unchanging variables.

So, I switched it this morning to use a singleton instance instead, so the class is only being created once and referenced thousands of times each import.  Much nicer already.

It's stuff like that that makes me wish I knew more about OOP.  I am studying it on and off, but there's still some concepts that I just can't wrap my brain around at times, like exceptions.  In my procedurally-attuned programming frame of mind, every time someone explains them to me, I think ... "Well, if something *breaks* why don't you just work with the return codes and work around that?"  So, yah.  Some stuff is still lost on me.  I'm trying to figure it out though.  Maybe it's one of those things that doesn't make sense so much when you apply it to PHP and it's general usage of websites.  A lot of the stuff I read about, I think how it would make much more sense if it were an actual application running.

Anyway.

On a totally different note, one thing I want to look at getting into the packages website is tracking a changelog of all the package's keywording history.  Right now, the import process is pretty simple -- if the content of the ebuild has changed, then the old one is marked for removal and an entirely new ebuild record is created in the database.  The reason for that is because that is far easier to do than it would be to examine all the myriad of data that is associated with one ebuild, track the changes, and then flag those.  Instead, I just dump the old one and treat the new one as a completely new record.

There's a tradeoff in the compromise, though, because instead of tracking ebuild modifications, I have to do all this coding to flag packages and ebuilds that things have changed and to treat them as an update instead of a new one.  That was tricky to get setup right, and getting that stuff in there in fact was one of the main things that pushed the initial launch back.  It was just one of those things that I couldn't run into the bugs until I started actually doing  a sequence of import runs, since they wouldn't show up until then anyway.

But, I'd like to start at least tracking the ebuild keyword status changes.  The reason is because that is really valuable data that can provide an excellent set of reports.  For instance, we can see which categories / packages / herds are getting ignored historically as far as stabilization.  Plus you can do cool stuff like import results from a statistics tracker as far as what people have installed, and you can start to see where maybe the tree could use a little more love.  And, it would help contributors who want to help out, but are overwhelmed by the enormity of bugs and packages and issues that need to be addressed.  I could see it being helpful saying, "here's an area that is suffering from neglect *and* is popular."  That would be cool.  And that's my goal.  In fact, that's *been* my goal for years.  I'm just now getting to the point where it's becoming possible, though.

Fun stuff.  I gotta hone my coding skills as I go, though.

I feel bad about how things have gone so terrible since the initial launch of the site. I really was not ready for the massive load, and my interim solutions were just slow and clunky. Hopefully things should be much happier now.

There's still a lot of silly bugs in the code that I need to fix. I just found another one this morning where the caching is breaking if you change your architecture selection around. Oy. I'd like to get to them, but I've been pretty swamped for time lately, between starting a new job this month and dividing my remaining time doing consulting work for two other companies.

Having a break from it though has been kind of good. I've already thought of a few optimizations that I can throw in there that are kind of like, "well, duh" type stuff I can't believe I didn't think of. For example, one way that I check to see if an ebuild is new is to see if the file mtime has changed. I don't know why it never occurred to me to just read the Manifest file and see if any of the hash sums have changed. That'd save me a lot of time.

I've been poked a few times about getting the code in a live repo somewhere, too. I guess that's coming soon, assuming I can get around to it. Personally, I don't like the idea of doing it when I *know* my code is in some ugly stages, but whatever. I need to learn how to setup a git repo anyway.

Oh yes, that reminds me. I also moved all the Planet Larry stuff onto the same server. Everytime I poke at the site, all I can think about is how much of an overhaul the whole thing needs. I'm totally embarrassed that I haven't even switched over to using Gravatar yet.

My goal is to ditch the planet software and write my own software to pull in the feeds, drop them in a database, and have the whole thing searchable. Then build a user admin section as well so users can manage their feeds themselves, and stop waiting on me. I'm planning on making that my next project, once Znurt gets to a better stage of stability.

Right now, though, I just did some minor tweaks. I got rid of the subdomains, and all the other projects on the site that I let atrophy, so planet is just available now at http://larrythecow.org/

El sitio esta aqui: http://es.znurt.org/

Translation: I'm all out of cookies.  May I date your daughter?

You may or may not know, but Znurt is the name of another Gentoo mascot of ours -- he's the little UFO / whatever flying around that you can see on Gentoo's homepage.

I'm not real sure what else to say about the site, other than it's been a lot of work, and I'm really proud of it.  If I had any idea of how many hours it'd take to put together, I never would have started on it, that's for sure.  I think it's safe to say I've put in at least 200 hours working on it in the past 2 months or so.  I think it's turned out really well, though.  It's a far better improvement over my old packages site which was thrown together quite hastily, and the code was pretty nasty.  This time around, everything is object-oriented which made it really go by quickly.

I've tried to put in some of the best features for launch as possible, but some just had to make the cut due to constraints.  The ones I really wanted was a page for new ebuilds, and a chance to have a feed of those, and customized by arch.  Same for new packages as well -- a view for each arch.  The new packages section almost got cut completely, though, because of some issues I ran into at the last minute, but fortunately that one made it.

I'll still be adding more stuff, but probably not at such a clipped pace as I have been for the past month.  You can track the planned features I already want to implement here.  If there's other ideas you have, let me know, and I'll see if I can get them in.

Which reminds me, I'd like to give a huge shout-out to all the beta-testers that helped out with the site development.  Every single person that got involved had a unique idea or perspective that I hadn't even considered, and it really helped clean things up quite a bit.  Thanks, guys, I owe you one.

I hope you all enjoy the new site.  Please contact me about any bugs you find -- hopefully I didn't miss any obvious glaring ones.  Also, if there are any XML wizards out there, I'd really appreciate some feedback on the Atom and RSS feeds, in case I missed something.  Not really a strong point of mine.

The thing I like the most about working on the site is it really revitalized my interest in Gentoo, and working on the project as a whole.  I had been losing interest in it for a while, but I realize now that that was partly fueled by my slip-shod attempts at user websites (planet included).  Taking a weak thing and making it strong, though, has really helped to push me in the opposite direction, though.  Not to say I'm going to make a career out of Gentoo development or anything, but it sure does help fuel my passion when you have some work you can be really proud of.

In fact, working on this has inspired me to do a rewrite of Planet Larry as well.  Writing the XML feeds wasn't nearly as difficult as I imagined, and though I don't understand them too well, it's definitely something I'd like to learn more about.  So, I'm going to ditch the planet software we are using and write my own, and finally give Planet Larry some features and love that it really needs.

Oh, also, there's a twitter feed that I use just for the website development, if you want to see what I'm working on, there.  Have fun.

The new design is in place, though, and it looks awesome, in my opinion.  I can't wait to show it off.   Right now, only the testers get to see it -- if you'd like to do some as well, just lemme know.  In the meantime, here's a little preview:

Now who is that little guy?

I've also got a Twitter feed set up now where I'm sending updates about the site progress, if you're interested in following.

I would cover a list of new features again, but this time I think it's embarassingly short since I've been just fixing bugs for the past few days.  I did get really basic RSS feeds added, but that's about it.

Here's some of the planned new features, though:

• display changelog with syntax highlighting
• display ebuild source inline
• show use flags, dependencies, reverse deps on ebuilds
• show open bugs on ebuilds (thanks to Mike for his help on this one)
• RSS feeds: new packages, version bumps
• Compact, verbose views
• Text-friendly design, for CLI browsers (elinks, etc.)

Like I said, no idea how long this stuff is gonna take, but it should make the site much more friendly and usable.  Yee-har.  Lemme know if there's something else you'd like to see, and I'll fit it in if I can, and it's reasonable.  Thanks, guys.

I have been working on this thing almost non-stop all week, because this site is by far the most complex one I have going for me.  Well, that's a personal project, at least.  I forgot how much work goes into this thing, as I've still got small laundry list of to do items.  But!  The good news is it should be ready and online really, really soon now.  Hopefully this weekend if I can squeeze more blood out of me.

I got to see the new design this morning, and I gotta say, I am absolutely floored by how amazing it is.  My brother-in-law was kind enough to do an original design for the rewrite, and I gotta say ... it's just spanky.  I love it.   He also redid the design for Planet Larry if you want an idea of his skill level.  Thanks, David!

Anyway, I'm not gonna open up the site just yet, but I do wanna start getting a list compiled of people who are interested in doing some beta testing for me.  Just send me an email if you'd like to help out.  It's not really a big deal if you wanna do it -- I just need you to poke around, use it like you normally would, but (and this is important) actually send me feedback about any bugs you find or suggestions you have.  Lemme know.

This is an incomplete list, and I've just started jotting it down yesterday afternoon as I started working on the site some more.  Also, it's a brain dump, so excuse the randomness.

Roadmap:

• package masks
• website; new design
• load testing
• bash script w/import options
• expanded versions
• options to use find all ebuilds updated recently
• status column for updates to happen in background
• db classes to access properties

I'll explain quickly what some of those mean.

Finding out whether a package is masked or not is a real pain in the butt.  The reason is you have so many ways a package can be masked.  I'm not gonna go into the coding required to checking it.  It's possible, and it's nice when it's done, but it's painful to write out.

New website -- yes, I'm rewriting it from scratch.  Well, the backend.  The functionality is going to be there 100%, so nothing is going to be lost.  If anything, there's gonna be a lot more stuff.  I'm also hoping to get a new design done before launching it.  Also, I'm probably going to do a closed testing invite session before launching it, so I can get some serious feedback first.

Okay, the status column thingie -- that's gonna eliminate one really annoying feature of the last website.  It would shut down for about 5 minutes while the site was updating, because it would delete stuff and add stuff all over the place.  The new one is just going to insert all the new stuff in the background, but won't flip it on to be actually visible until the import is completed.  So, it'll be completely transparent to the user, and the site will always have populated data, and any updates will just show up all of a sudden.

Now, some of the feature requests:

• RSS feeds
• XML API
• import all metadata
• parse changelogs
• profile masks
• tags
• track deleted ebuilds, packages
• screenshots
• user ratings
• twitter feed
• GLSA integration

A quick disclaimer -- no idea if and/or when any of those are going to be done, because they are all unnecessary to the actual launch of the site itself, which is what I'm working on now.  Another quick recap on what I'm planning, though.

The RSS feeds are going to be much better this time, and I'll have a larger selection: new ebuilds, recently updated ebuilds, per-arch feeds, etc.  And they'll have all the same info on the website too.

The XML API, which I've mentioned before, isn't really going to be that fancy to start with.  You'll just be able to do something like browse to website/app-admin/foobar/xml and it will have an XML printout of all the data that I have on that package.  Same for category pages, too.  Nothing fancy to start with.  Oh yes, I'll also produce database dumps as usual.

I need to import all metadata too to get GPNL back on its feet.  I already *have* it all in the database, but just as the raw original strings.  I need to sort through it and get it into its individual tables.  Again, I'm just doing the bare minimum right now to get the site up and running.

Parsing the changelogs, ugh.  That's a new feature I tried adding in the old website, and it never quite worked out exactly how I planned.  I'll get that one in eventually.

Profile support is one feature I'm really excited about, and one I wanted to keep a lid on for as long as possible.  Oh well.  Now you know.   Basically, I'm going to add support to the site to browse the keyword statuses for a specific profile instead of just the default one.  Once you change your profile, it'll affect package masks ... and, that's about it for now.  Still I think it'd be a handy feature.

Tags, screenshots, user ratings -- just some features that I'll eventually add to make the site actually accept user input.  That's gonna be a long way off because of all the work involved in user accounts and preferences and uploads.

Tracking deleted ebuilds -- I'm still debating whether I want to do this one or not.  I mean, I'll track them in the database, but it could be inaccurate for a number of reasons.  Not sure if I'm going to have a display for it on the website or not.  Would be kinda nice, though.

GLSA integration will hopefully be easy enough, I haven't looked at it at all.   Just track what GLSA notices there have been for a certain package and display them when you view it.  Nothing really fancy.

Alright, so that's about it.  I know I've gotten some feature requests from people, and if you don't see it listed here, then I've forgotten about it.  Please do me a favor and ping me on IRC or email and let me know, and I'll get it added to the list.

I still don't have a timeline to get the old site back up and running.  The good news is it is using a lot less CPU than I thought it'd be, so that expands my options for hosting.  The coding for the original roadmap is coming along at a clipped pace.  Everything I need to do has either already been done and just needs to be rewritten for the new backend, or is possible without too much problem.  In other words, there are no major roadblocks.  I still see it taking a few weeks, though.

I'll keep you posted.

I've been working on the packages website rewrite over the holidays, and it is really starting to come together.  I've decided to short circuit the process and instead of going for a full-featured site that I'd like to have, just duplicate the bare minimum of the old site to get it up and running as fast as possible.

I don't want to raise hopes too much, but it's gonna be a really cool site -- not the first launch, but what it'll eventually become.  I've rewritten all the code to access the portage tree, and I have classes to access just about everything.  Doing that has given me a tremendous amount of flexibility, and it just makes things easier to bang out.

There have been some rough spots that I have run into that are some hurdles.  Right now I'm working on determining masked packages.  Always a bit tricky, but thank goodness that's the hardest part I'm running into right now.

I don't have a timeline for when I expect the site to be up.  I'm actually ready to start rewriting the frontend for it right now.  I'm looking at getting a new design theme for it, and I'll probably wait to get that before launching it.  I already have a new domain for it, too.  Much shorter, easier to remember and type.  Realistically speaking, I think the site is probably going to be going live somewhere near the end of December.

Hosting is still a problem.  I don't know what I'm going to do with it, partially because I'm not sure how much CPU or traffic load it's going to generate.  I have three options right now -- host it at home, on my Linode or on a shared hosting account -- neither of which I'm really excited about since I think they'll all have different things they'll choke on, network traffic, CPU usage and speed, respectively.  So it's gonna be very likely that it gets bumped around a bit at first.

One last thing I wanted to mention.  I almost decided to completely abandon the project completely, since it seemed like no one really cared that I had even taken it offline.  I have had a few people poke me though, and ask about it's status, and I've seen it mentioned in a few other places.  This is one of those projects that I'm never sure if anyone ever finds it useful, so lack of feedback makes me want to drop it completely.  Inversely, any feedback makes me want to work on it and keep it going.  So, I'm really curious to hear what you liked about the old site.

Also, along that same vein, something else I'm going to do with the new site is build out features based on requests only.  I have a lot of ideas that I'd like to put in there, but instead of doing that, I'm going to hold off on them and just build out what the users want.  Everytime someone tells me how they use it, it surprises me because I never imagined it being accessed in that way.  So, again, feedback is critical.

Well, that's all for now.  Again, I don't know how soon I'll have a super slick website up.  Chances are, that I'll put an XML API up first (another new feature) or the RSS feeds since that doesn't require any nice eye candy.  We'll see.

The new site will also run on my dedicated Linode, where I think I've finally correctly managed the apache issues, so that means there will be less arbitrary downtime as I screw around with my box here at home.  I really hate running servers at home that other people are dependent on, because I like the freedom to change things around without affecting anyone.  Right now, the old site is so CPU intensive, that I can't move it over to the VPS.

The code for the new site is much cleaner.  The entire thing is rewritten in OOP classes to access the portage tree, which makes my job incredibly easier.  Not to mention it's a lot faster.  It'll still be a bit before I get it online, but killing it will inspire me to push it along.  I'm tired of having this thing limp around when it's just a dead albatross around my neck right now.  So, farewell.  The new one will be better.

I've been ripping my TV shows, and so far I haven't seen any really hard and fast rules on what to expect with them on DVD.   Part of the reason is that I just haven't been paying much attention to subtitles until recently.

I was playing with ripping one show last night, and I saw the CC logo on the back of the case, so I went to check the rest of my library to see which other ones had it.  Nearly my entire library of Warner Bros. DVDs displayed the logo -- even for much older cartoons (Looney Tunes, Scooby Doo) -- once again staying consistent with the fact that the studio puts a lot of effort into the quality of their releases.

I just started playing with extracting CC though, and just barely wrote the code to my DVD ripper to extract them, so I have no idea what the other series are like, if they have subtitles or not -- VobSub or CC.  I usually don't find out until I actually go to rip them.

Extracting the closed captioning subtitles is a lot easier and faster than getting the VobSub streams.  For Linux (and Mac and Windows) there's a nifty OSS program called ccextractor.  Once you have your VOB video file on your harddrive, just run that on the movie, and it will create an SRT subtitle file of the closed captioning text.  It's great, and really fast, taking probably under a minute on a 60-minute video on my box.  Comparatively, when ripping a VobSub stream, you need to read the DVD directly which causes its own bottleneck, and then demux the entire stream.  It takes probably around 3 to 5 minutes for an episode of the same length.

Another thing I like about the closed captioning titles is that because they are extracted as SRT, it's easy to look through them since they are just text files.  If you're really anal, you can correct typos yourself.  The VobSub subtitles are all bitmaps.  I've also noticed that on some DVDs, where there were issues with framerates or something else, that the VobSub timestamps will be off ... and sometimes either they will show up clumped together at the beginning of the film or the sync will be way off.  I think that this has to do with the dumping process, somewhere, but I'm not sure.  I've never really taken the time to pin down the source.

So, with closed captioning being easier and faster to extract, as well as editable and the timestamps haven't had any issues for me (yet), it's quickly becoming my preferred subtitle format.

There's only one small issue with using ccextractor, and that is you won't know if there are any captions in the VOB until after it's made its trial run.  The program will create an .srt file regardless when you run it, but the file will be empty if it couldn't find any.  That's the only drawback.  With VobSub, you can know if there are subtitles just by probing the DVD using lsdvd or something similar.

Muxing it into matroska is simple, too.  Just pass it as a file argument and you're done.

As a sidenote, while my bend application that I wrote and use to rip DVDs would be a major pain to setup for someone else, I've rewritten it recently so that it uses individual classes to access every object directly: DVD, DVD track, DVD VOB, Matroska file.  They are standalone classes written in PHP if anyone wanted to use them, feel free.  You would also need my tiny class of shell functions as well, since they all make calls to it.

The DVDVOB one makes it simple to extract the subtitle stream.  In fact, all the classes make things relatively simple.  They have made writing my code so much simpler.

The first step, though, was getting the PS3 to run Linux.  I took the shamelessly easy way out (and I don't regret it either) and installed Xubuntu.  I won't go into details about how I got Linux on my PS3 since that's well documented.  I will say that I remember quite vividly now why I can't stand binary distros.  Bleh.

The BD filesystem is UDF.  Providing you have a recent kernel (2.6.20, I think) with UDF v2.5 support, you are good to go.  I mounted a remote share and just dumped the disc to an ISO file onto my desktop.

$ cat /media/cdrom0 > wonka.iso

That was the easy part.

The hard part was trying to get it decrypted.  I had to use Java tools (bleh) to get to the source.  There are three applications you need.  And if you hate digging through forums and using download services, then I've got direct links for yah:

• aacskeys-0.4.0c.tar.gz
• bdvmdbg-0.1.5.7z
• dumphd-0.61.tar.gz

For Gentoo, you'll need to install the JDK to build the aacskeys library and binary.  I just emerged dev-java/sun-jdk and it worked for me (I know absolutely nothing about Java, but my stabbing in the dark miraculously worked).  You'll also need a runtime environment to actually execute the stuff, and I emerged dev-java/sun-jre-bin and that worked fine, too on my amd64 box.

For aacskeys and Gentoo, you'll need to apply this patch that I cobbled together from what I found on the doom9 forums to get it to compile.  It just fixes the Java include directorys for the Makefile.

Now, I'm still a bit fuzzy about what each program does, and whether you need all of them or not, so I won't go into a lot of detail.  What you want to use, though, is the dumphd program.  But to use it, you'll need to copy the aacskeys library and a file from the bdvmdbg package as well into the path or same directory as the dumphd program.

Once you have that, you can just run dumphd.sh and it'll fire up a simple little GUI telling you if it has all the libraries it needs.  Then you just specify the source and destination, and aacskeys will see if it has a working key to access the disc.

I can't really give much more detail than that, since I'm so new to this.  Suffice it to say, if you read the accompanying README doc that comes with each one, you'll get along just fine.

It took me a long time last night to get just one disc ripped and transferred over my subnet to try it out, and by the time I managed to get it mounted (mount -o loop -t udf wonka.iso /mnt/udf) and access it, it was pretty late.  The keys I had didn't work for my disc, and I didn't want to try the whole procedure over to try another disc.

Anyway, good luck if you try it.  One thing that impressed me is how much simpler it was than I thought it'd be, but what a pain it was trying to figure out where things went wrong.  The doom9 forums are a good resource, but not exactly the best place to find clear, concise information for a beginner.  That part was frustrating.

Personally, I don't think it's worth the hassle right now, the way I did it.  I'll get a BD-ROM sooner or later so I don't have to transfer the content over the network and can instead just test it directly.  But, I started out to see if I could at least get a copy of the ISO and get the tools running all without Windows, and I can.  So, that's progress right there.

I know I've played with Feedburner in the past, and kind of flip-flopped on whether to use it or not, but this time I'm sure I'm gonna stick with it.  It's better for users, since they will always have a feed available (whether I have issues or not), and it's better for me since I can offload that part of the network traffic, which is actually quite a lot.

I've already updated the feeds and my apache config to do a permanent redirect, but if you want the feed URLs directly, here they are:

• Planet Larry: http://feeds.feedburner.com/planet_larry
• Larry the Universe: http://feeds.feedburner.com/universe_larry

Sorry for the inconvenience.  It seems like everytime I post about Planet it's bad news or maintenance.  Believe me when I say that it aggravates me far more than it does you.

Specifically, the issues I'm having is that Apache is sucking up all the available RAM, of which I only have 360 megs on my account.  It's then rolling over to using all the swap space as well, which only slows things down even more.  I've just started playing with tweaking the MPM configuration a bit, and I'm still trying to find a reasonable solution for my configuration.

In the past, the Linode had been seizing up occassionally, and I'd normally just reboot it and get on with my life.  Recently, I installed monit (an awesome app), and pinpointed that the issue seems to always be with apache.  Now, I'm just trying to narrow it down even  more from there, but offloading the RSS feeds seems like a good step to take anyway ... I get gigabytes of traffic per month just on that, believe it or not.

I'm toying with the idea of setting up lighthttpd instead, but I really prefer apache, and would rather set it up to behave in a low memory environment instead.  So, for any downtime in the near future, chances are it's just me tweaking something.  At least now, thanks to monit, I have a much better idea of when something goes wrong.

Oh, one other tweak I've made is that the planet script itself is more robust as well.  That thing used to run out of control, but I've made some changes that will ensure that if it runs away, at least it won't bring down the system.  I also started playing around with the idea of writing my own feed parser to replace the Planet software completely, and it looks like it's going to be much simpler than I imagine.  I haven't actually started down that path yet, since I have bigger projects to complete, but I'm actually enthusiastic that it'd be far, far simpler than I imagined.

Please be sure to hold your applause until the end.

Acutally, to be honest, I'm a bit nervous, since I have no clue what I'm going to say, and I don't do well with crowds.  Should be interesting.  Maybe it'd be a good time to pitch Ubeantoo.

There's a couple of reasons for this, but the most important is that it will give us flexibility to deal with changes from upstream.  For example, with 1.8.1 and above, PEAR changed it's base XML dependency to XML_Util.  That one was already in the tree, and so the new pear base system relies on that.  If the other base ones change between versions, we can focus on that.

Another nice little change is that the base system ebuild now is just dev-php/pear.  So, "emerge pear" and you're done.

The new versions are all currently marked as unstable across the arches.  I would appreciate any and all feedback on the change.  I'm still a bit skeptical that this is the best approach, and a bit nervous at any fallout that may occur, so please file bug reports and let me know if you have any issues.  Thanks, all.

There are two *places* to get the drivers from: either in the kernel, or from the alsa-driver package.  But, when using the kernel drivers, like many other drivers there, they can either be compiled in statically or loaded as modules as the computer is booting up.  So, there are actually two ways to install the drivers as modules, which could be a bit confusing.

So, a quick list:

1) In-kernel drivers (statically compiled)

2) In-kernel drivers (modules)

3) External drivers (alsa-driver package, modules)

The first two are the officially supported methods by the ALSA team, so I'll quickly focus on those two.  Now the, recommended way to do things is #2 -- select them as modules in the kernel and build them that way when you are setting up ALSA for the very first time.  Why?  Well, the answer is really that it gives you a lot more options.

Let's say, for instance, that you aren't sure which driver your card requires.  So, you flip on a few that look like it's the right one, and set them to be installed as modules.  Once they are there,  you can run alsaconf, which is a part of alsa-utils.  The alsaconf program will do the detective work for you by looking at the modules that are available on the system, and the cards that you have on your box, and then load the modules and update your module list so that they will load up the next time you boot your box as well.  Pretty simple, right?  It sure is a lot faster than compiling one driver in the kernel, rebooting, testing if that works, trying a separate one, rebooting, etc.

Another reason is that there may be some options you need to pass to your module.  This is rare, but it does happen.  If you are loading them as modules already, then it's just a simple tweak to do change the settings, again, without having to reboot and re-test everything.

So, that's the reason we recommend you load them as modules.  It's just gonna make life a bit easier the first time around, as you are trying to determine what you have.  Once you know what driver is required, you can always go back into the kernel and compile it in statically, and be done with it.  There's no reason to keep it as a module, unless you want to.

Finally, a quick note about the alsa-driver package.  It's often said that it is unmaintained, and the reason for that is because I, personally, am the only one who is keeping it on life support.  That is, I'm the maintainer, not the ALSA herd.  It's only in the tree as a convenience to people who need to use it for whatever reason.  Some of the reasons could be that you needed to see if the latest release from upstream is fixing some issues of yours, so you'd use the live ebuild.  Or, you may want to use an older kernel but still keep the newer version of ALSA.  Or whatever.  The problem, though, is that I don't have the technical skills to troubleshoot your issues if something goes wrong.  My solution every time is  pretty much going to shrug and say "Sorry, that sucks.  Try the live ebuild, or something else."  It's not that I don't want to help, it's that in this case, I can't.

Anyway, that's it ... I hope that clears up a few issues.  When I have time, I'll be revising the ALSA docs.  No idea when that'll be though.  Don't hold your breath.  In the meantime, if you have issues, my recommendation is to post on the Gentoo Forums in the Multimedia forum and ask for some help, or there's always bugzilla.  Chances are you'll get a response faster on the forums, though.  Good luck, and God speed.

Looking at the detailed changelog, there are a lot of updates for the *hda cards, which is just what the world needed.  I haven't ever had any problems with them myself, but they are the de facto chipset on desktop motherboards right now, and it's awesome to see support getting improved.  Hopefully it'll fix some of the countless issues Gentoo users are experiencing.

And for the record, here is my first line of defense that I will tell anyone when they have no sound / wrong driver issues:

- Use latest ALSA (and kernel) in the tree. In this case, it's 1.0.20, and we've cleaned up a lot of crap that was causing issues.

- Compile everything sound-related in the kernel as a module. Yes, everything. No, not just the stuff that you *think* you can statically compile in --- *ev-e-ry-thing.* Once you've done that, run alsaconf. If your card doesn't show up, it's either completely unsupported (unlikely) or you didn't pick the right drivers. Go back in and flip on more stuff. As modules.

- If the correct drivers are loaded (aplay -L works), and you have no sound, then your mixer levels are probably wrong.

Apologies for coming across so heavy-handed, but that's the approach that has always worked for me, and works for most everyone when I tell them to do that: latest kernel, latest ALSA, everything as modules, alsaconf, then mixer.

Edit: Just as a postscript, the reason for the bluntness is that I recently started taking a more active role in the ALSA herd. Normally, I would just do version bumps and leave it at that. But, I'm trying to get practical issues resolved, and the one common thread I see over and over and over again on the forums and IRC is people just going off on their own little path and stabbing in the dark to see if they can get their sound to work, and then acting surprised that some random configuration didn't work.

I want to very much find both the source of this disinformation and correct it, and make sure that setting up sound is a really simple process. The *problem* though, is that it's always been pretty extremely easy for me, and so I am having a really difficult time understanding *why* people are hitting these issues -- hence, the reason for the forum post asking people why they are running into problems, and what thoughts and conclusions lead to that direction. Once I figure that out, then I can go back and tweak the docs to clarify situations and attack the common confusion.

I keep getting the idea that we could use a really simplified version of the ALSA doc, one that is a quick basic howto get setup for people with one soundcard, and then include examples using the most popular hardware out there right now (intel-hda). Then, have a separate doc for more advanced issues, for people who need help and information on alsa plugins, or have multiple sound cards and things like that. I think that lumping them together into one, as it is now, makes it hard for people who just want to get up and running.

So, apologies if it feels like I'm browbeating here, but the reality is that I'm really motivated to streamline the process of getting sound in Linux. Nothing but love, yo.