Recent blog entries for Svartalf

Wow... 2007... That's a REALLY long time to be putting off blogging. So what have I been up to...well...

Got several differing contract jobs, including with EF Johnson and Tektronix Texas, LLC, now in Plano. I'm currently working in the Boulder, Colorado area for Covidien doing systems design consulting work.

Got several non-LGP porting contracts. Kind of drifted away from them, but I've not totally broken off the conversation with them (even if they think I might've done so...)- I'd like to think I might bring something to the table for them that'll put wind back in their sails in spades. Caster's been a fun project as has Cortex Command- and being one of the devs that helped put over a half million into the hands of the EFF and ChildsPlay (Something I'm quite proud of above and beyond the games themselves...). As a result of those two games, I'm in the process of refining the toolchain I've come up with that allows me to grind out titles for X86, X86_64, ARM, and whatever else might come after- and in an easy to work with manner for those not steeped in Linux.

This brings me to the cause for me to come back over here...to vent a bit.

I'm in the process of bringing in glib temporarily into a toolchain sandbox, which, while it's part of the GTK project, is often times used as a generic toolbox that has NO relation to UI code. I find that the idiot scripting for GLib insists upon having gtk-docs and the absolutely latest libtool...so far

I can almost be made to understand the latter, but not the former. Optional item, yes. Required for building the library? C'mon guys... There's projects and conditions that do not and can't use GTK+. GLib's baseline and used for a LOT more than the UI toolbox- and should probably have, if you need documentation, a PDF/PS file somewhere in a ./documentation directory. Just because they're used by your UI project doesn't mean I want/need your documentation for the rest of the framework- or honestly need it installed...unless I'm doing GTK+ related work. Separate items...separate documentation with the online docs being OPTIONAL for any of it.

17 May 2007 (updated 5 Aug 2007 at 04:22 UTC) »

Well, long, long time since I posted a blog entry... Fitting now that I'm getting home, working on getting a new day job gig (Heh... The work for AMD didn't pan out QUITE as planned... I was led to believe I would be eventually working on Linux support with the Orca project codebase for the ATI card drivers- that didn't even come close to happening... I'm going to refuse to broach my NDA, but I will say I hope they fulfill their commitment to the Linux community- but don't expect anything good to come from that for at least 12-24 months in the future...) But hey, can't have everything.

Ballistics:

Ballistics has gone GOLD- WoooHooo!!

Expect the official announcement from Michael Simms shortly, but it's out the door and I'm hopeful we'll see Bandits: Phoenix Rising and Disciples 2 out the door fairly quickly here.

Anyhow, I'm out of here for now- have fun folks!

Update on the white/yellow gold "story". Seems that Helzberg did an even 'cuter' thing with my wedding band... Instead of yellow gold accent inlays against a white gold (albeit grade 3) band, they took and used yellow gold on the crown, soldered a grade 3 white piece for the bottom where it'd normally see wear so it'd hold up better as a "white gold" piece and then Rhodium plated the pieces that were supposed to be "white gold". The reason why I say this is that upon closer inspection with magnification, it's very obvious now that it's just plated yellow gold on the crown and not white. Gotta wonder what's going through the discount jeweler's minds when they advertise this stuff- it's definitely a questionable practice (I'm not quite sure it's fraudulent, but it's awfully close if it's not...) to represent the work as "white gold" when it's like this. It's not THAT much more expensive materials cost wise to do this the "right" way- why be shaving a couple of dollars off of something that will definitely be found out and make for a lost customer that will tell other people?

If you're looking for a white gold piece of jewelry, one should watch out for the stuff- according to Manufacturing Jewelers and Suppliers of America site, there's three grades of the stuff out there, each grade happens to be yellower than the previous. If you can find grade one jewelry (Not going to be cheap or easy- and it likely won't be from some place like Helzberg Diamonds...) you'll be okay as they don't end up Rhodium plating the stuff. Most of what passes for "white" gold these days apparently is of the low end of the grade two or solidly in the grade three slot. And they have the unmitigated GALL to charge larger sums of money for "white gold" when in reality, they're selling paled yellow gold with Rhodium plating. If you've not figured it out, my wife bought a white gold wedding band back five years ago for my Anniversary present, to replace my original band. The crown of the band has a distinctive yellow glint (Grade 3 "white" gold) from wear and tear. Basically, it was a poorer grade of jewelry and she paid quite a premium for it from...Helzberg. No, I'm not going to say they "gyped" her- a lot of their other things are quite nice and worth the price- just NOT the stuff they're selling as white gold. I bought a replacement wedding ring for Her for her birthday (When I had money...) last December, from Helzberg and they gave me the story about the Rhodium plating and that all white gold does this so I've been wondering since that time. Well, that's flatly untrue- it depends on what they use to kill the color. Traditionally, it was Platinum that was used to make white gold- you don't need a Rhodium plating on that stuff unless you want it really bright and shiny and don't want to bother with the polishing effort on the work to get it there. However, they also use stuff like Nickel (which is where people come up with stories about "White Gold Allergies" and the yellowing comes from- it's a cheap replacement for the noble metal whitening agents, but it needs a Rhodium plating of some sort to make it that brilliant white we associate with White Gold. It's the cheap stuff, and many of the jewelers just don't have the guts to admit that while it IS still gold, and it is sort of white, that it's just not the White Gold that your mother had.

6 Jun 2006 (updated 6 Jun 2006 at 04:36 UTC) »

Well, It's the beginning of another month. Things are still a little tight, but the light at the end of the tunnel ISN'T an oncoming train.

Ballistics:

One bug left. Not a show-stopper by any means, but we'd like to get that one cleaned before going gold- that insistence on quality sort of thing (Especially since it took as long as it has to get out the door...). It's nothing that is going to have us NOT ship at this point, but we're being picky at this point about it. Michael's got another nice game that is about to go out the door (Many will want this one- honestly... It's dead cool.) so there's a couple of weeks delay before we get serious about taking it to the duplicators for release- so I'm taking the time to zoom the last known issue and to evaluate the possibilities for PowerPC and AMD64 ports. They probably won't be included in the main CD and would be a follow-on installer of some sort- don't know how Michael will want to do it, but if we get a port going, I'll announce it here and probably in a few of the other venues.

It's nearly there, gang, honest. The upshot of all this is that Bandits WON'T take anywhere near as long, and we ended up with some tools that will be properly Open Sourced once we get them much closer to something usable by someone other than LGP.

NetEffect:

Not much to say (Covered by NDA, afterall... :-) other than it's working at paying the bills incurred over the period resulting from Nexa leaving me high and dry. They're tickled with me as best as I can tell- I'm getting access to hardware to test the stuff I've been working on tomorrow sometime. Three weeks, and I'm already where they expected me to be in about two months' time- that's not an assessment of what they thought of my abilities, that was an honest assessment of the amount of work. Doing 10-12 hour days (Overtime on contract pay is a good motivator, no?) will do that for you. I've got about another 2-3 weeks or so of the code crush before it lightens up a little for me and I get it code complete for the current hardware. Besides, they're making things I'll need if Coollogic gets it's funding and I'm getting paid to write code that I'd have had to contract out for if they weren't doing it right now- not too bad. (Did I say "if" there? I really meant "WHEN".)

Coollogic:

Major international contract that is expected to be closing all the way tomorrow. Can't say more on that one, but the work really begins closer to the ending point of the NetEffect work. We'll need interconnect hardware for the aggregation clusters, etc. NetEffect, if their silicon's decently stable (and it looks to be so far...) it's the front-runner for that right now. Funding is purportedly about to start the final stages at any moment- and I've seen enough to believe that this is very likely to be the case.

I'll get excited when I see that first draw-down in the operating account and I've drawn part of my back pay, but it's beginning to actually look good- and not in the wishful thinking sort of way. I just wish I'd not paid 5+ years of pain to get there.

20 May 2006 (updated 6 Jun 2006 at 15:59 UTC) »

Well, it's May... Soon to be June.

Ballistics:

Scratchbox WORKS well. We've got a version of our code that compiles clean, runs clean on Debian Woody and above (i.e. glibc 2.2 and above...). But... Well, I can't divulge what's the holdup this time, but we're only a week or so from gold (crossing fingers) as we've got only a few items left, and none of them are real show-stoppers. The delay's due to something cool that'll make everybody's lives just a little better in the game-dev industry. (We hope...) I'm waiting for the go-ahead from Michael to merge the non-network fixes I've done back into the main branch- while I'm waiting, I'm supposed to be vetting the new build environment I've cobbled together, and kick-starting Bandits and Disciples 2. I WAS going to do Bandits, but, because I've only my laptop to work with, it's going to end up being Disciples 2 instead, I think. (I'll explain that statement in a moment...)

NetEffect:

I've taken on a short-term contract doing Linux related support for one of the main iWARP players to help make ends meet until things gel with Coollogic (And they are going to, it's just taking longer than we'd envisioned- way longer than anyone had envisioned, in reality...). 10Gbit Ethernet. TCP Offload Engines. High Performance Computing. I get to play in that playground and get paid (well, at that) doing it. As a result, I'm down in Austin, TX for the next couple of months- so I'm only at my main setup for a limited amount of time (a couple of weekends a month...). This means I've got to use my lappy for the development of things after hours, so no high-end 3D because ATI's Xpress 200m doesn't perform all that well (At least 50% slower than an 9200 on an AMD64 machine in 64-bit mode as best as I can tell- it does about 10 or so FPS on Unreal engine games like UT 2k4...) because it's limited to UMA operation only right at the moment. Nice. Anyhow, I should be able to get a start on Disciples 2 once work settles in a bit (got a first week fire-drill...yeesh...).

I take it back... The month's been waaay problematic, but not QUITE a wash. I just got something going with Ballistics.

Ballistics:

Yow. Like night n' day, it is. I've got some cleanups to do, but it's now largely working like it's supposed to be. What did I use to accomplish this? Scratchbox and elbow grease. Didn't get to do Hyena's this Friday, but the end result of that was that I got to the place I did today. Either we will use my fork of the codebase or the one they've been working on up to this point in parallel with me- but we're almost there. This means that I'm going to be starting the code work on two other games. Bandits: Phoenix Rising and Disciples 2. Should be short work on both of them- magic words there, mind.

Well, this month's been a wash...
No day job (got laid off for "financial" reasons...hmph...), Ballistics is still being a pain (almost there, though...), ended up with what is very likely to be a Rogue Outlaw with one of the horses I let my wife buy. Nice. I can only hope that I'll get a new day job while I wait on the close of funding with Coollogic, I'll have another great night at Hyena's (so I can forget this BS for a little while at least...), etc.

Just got in from going to Hyena's over in Arlington, TX. (One of the main Comedy Nightclubs in the area...)- by and far the best experience I've had seeing comedians to date. Chad Miller opened up the night- calling him VERY funny man would be a major understatement. The Headliner was none other than John Bizarre; damn, that man is hillarious. I don't think I laughed as hard in ages as I did in that show tonight. If it ever gets back to you guys, again, THANKS for the laughs!

Ballistics

UGH...

Different problems (i.e. It's segfaulting in completely other areas when it's compiled with 3.4.5- and GDB appears to be less than helpful because it's handing back nonsense. I've seen this before in the past and even things like TotalView couldn't tell me much of anything useful- which is problematic... Had to debug things the old-fashioned way, with printf's the last time I had problems like this one.). Still looking for what gives with all of this. I'm still hopeful for an end-of-month solution, but it may be middle of February before we're into final beta at the rate things are progressing.

Ugh... What fun...

GCC, of late, is something of a pain for cross-compile people. Later versions (3.4 and above) strictly enforce the rule about -I include path inclusion. If it detects that it's something that "duplicates" one of the directories (okay, HOW do you guys "KNOW" that it's duplicating it??) it ignores it outright. If you do an -isystem to override some of the problem -I includes, it works just fine (okay, why add a new flippin' flag?), but you've got to be careful what you include via -isystem as some of the paths won't be found correctly (What? You mean to tell me that this path's not found even though I can 'cd' to it?) and you have to guess at which it will and won't balk at. I finally got the cross-compile environment to work to compile Ballistics working without resorting to Scratchbox (which is a little more cumbersome than what we're doing with LGP's setup.)- which is a win. Now, to see if I can get away with 3.4.5 or if I need to do 4.0.2 or 4.1 to build Ballistics.

I thought this was all about programmer responsibility and choice- not the developer of the tool knows best. That's what MS does, for goodness' sake- if I override the search order with a -I, that means just THAT. If it causes problems with things, that's MY problem, not yours.

44 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!