<terpstra> #scala
<terpstra> ups =)
<kristianpaul> ha, interesting http://www.ohwr.org/projects/wishbone-gen/wiki/Wbgen2_ram it seems can handle clock domains for me :)
<kristianpaul> hey terpstra ! :)
<terpstra> hi
<lekernel> hi terpstra, welcome back :)
<lekernel> were you at icalepcs?
<terpstra> yep
<terpstra> but not for the workshops on sunday
<lekernel> I also hanged around a bit on the next days, but I missed you...
<lekernel> the industrial exhibition was interesting :)
<terpstra> did you see the workshop on DSLs ?
<terpstra> tutorial*
<lekernel> I learned about ADCs, diamond-based neutron detectors, high precision piezo drives and 1.2nm position encoders
<lekernel> no
<lekernel> did you?
<terpstra> yes, i'm learning MPS and scala as we speak
<lekernel> "Scala is fully interoperable with Java."
<lekernel> hum :-)
<lekernel> is that a selling point? *g*
<terpstra> it depends what you want
<terpstra> scala is about as high level as languages get
<terpstra> yes
<lekernel> also java-based... hum
<terpstra> no
<terpstra> MPS's GUI is java based
<terpstra> the code it generates can be C
<terpstra> or something else
<terpstra> the cool thing about MPS is that you aren't editing one language or even text
<terpstra> they call it a "projective editor"
<terpstra> which means it represents its AST as text which you type as normal, but really you are operating on the AST
<terpstra> that means your code can include tables, flow charts, ambiguous grammer, etc -- and it all works
<lekernel> is it a bit like lisp macros?
<terpstra> no
<terpstra> this "projective editor" concept is really something new
<terpstra> scala was also quite popular at icaleps... it supposedly unifies OOP and FP. as a FP programmer who has to work with OOP folks, I am seeing if it is a nice middle ground =D
<lekernel> looks like a lot of clicking
<wpwrak> compromise-driven programming ? :)
<terpstra> excellent talk!
<lekernel> what's so great about Java? frankly (and sorry for that) I would describe it as some over-marketed junk that is merely about combining the disadvantages of compiled and interpreted languages
<lekernel> so I'm a little dubious when I see programming language talks from people who enjoy Java :) that's just a superficial first reaction of course
<terpstra> i have never liked java
<terpstra> but i'm not a bigot either :-P
<lekernel> grmbl. recording v4l video sources is amazingly complicated
<lekernel> the GUI programs are riddled with bugs and the command line is a mess
<roh> lekernel: well.. i simply have the pov: 'my life is interresting enough to be bored by java' ;)
<roh> about v4l, make sure you read recent specs, not v4l1 from the stoneages.... v4l2 cleaned some things up and added compressed video formats
<lekernel> well, I finally got it to work, with a multi-line gstreamer command ...
<roh> ;)
<roh> yeah. but in the end it IS kinda lots of details neccessary
<lekernel> now I understand a bit better why so many artists have Macs :-)
<roh> that doesnt help much ;) believe me.. i work in a house full of artists
<lekernel> at least connecting a TV card and capturing a video file (including synchronized audio) works out of the box, no?
<lekernel> without kilometer-long commands, GUI programs crashing, etc.
<roh> no.
<roh> only if you buy a expensive 'made for mac' product. which is like .. ever 100st tv grabber
<roh> so basically you buy more support upfront (if you get it or not) .. the products IS more expensive than the 99 other grabbes which work with win32 and linux
<roh> eh are
<lekernel> you have to take into account your time investment, I have wasted some 3 hours already on this
<roh> lekernel: also.. mac products only do ONE usecase.
<roh> if you try something else you are completely on your own, like on win32 or linux.
<roh> my guess is: you try somthing where there is no mac product
<roh> after all.. who does videocapturing nowadays? its all digital or has hw compressors
<roh> only geeks and professionals (who know what they are doing and use professional equipment)
<lekernel> I wouldn't expect a DV/Firewire camera to work any better on Linux :-)
<roh> heh.. you would be surprised
<roh> i tried getting one to work with some win32 tool from adobe once... didnt work
<lekernel> the mess I've been through is mostly because of userspace bugs and shortcomings
<roh> plugged it into a linux pc and installed kino, klicked 'connect' and i could remore control and copy stuff of the tape and back like i wished immediately
<roh> sure.. there are expensive mac tools i guess who do that. but i dont have them, i dont want to pay for them and i for sure do not want to select tools by vendors ideologic noncompatibilites anymore.
<roh> lekernel: heh. did you write up which ones? blog about such stuff (in a constructive manner).. or write up bugreports. atleast we can help make it better
<mwalle> lekernel: will the dvi out use the IOs which are currently available at the expansion headers?
<lekernel> mwalle, probably not, we need those special I/Os that support TMDS and they're on a different bank IIRC
<mwalle> TMDS? differential ios?
<lekernel> yes
<lekernel> for DVI-D
<mwalle> i guess they are unrouted on rc0 :)
<lekernel> they're unrouted on all boards
<roh> nice. sure you can use them without line drivers?
<lekernel> roh, seems so
<roh> we should do some tests to make sure before rc4
<lekernel> seems pretty difficult as those pins are buried under the BGA :)
<roh> in the end... people connect really long cables with a lot of capacitive load to that
<lekernel> we'll just check datasheets
<roh> hm.. do you have a spartan develboard which has them routed?
<lekernel> no
<roh> do you know of one? maybe i can borrow the right board
<lekernel> I don't even have any screen that supports DVI or HDMI
<stekern> roh: I've got a (spartan6) devboard with and without line drivers, a guy with a similar board did some tests comparing them, he didn't notice any big difference
<lekernel> there you go :) thanks stekern
<lekernel> stekern, got the schematics of the devboard without the line drivers?
<stekern> lekernel: sure, wait I'll dig them up
<stekern> meanwhile, here is the post where he is speaking about his test: http://rubidium.dyndns.org/pipermail/fpga-synth/2011-April/001667.html
<lekernel> thx
<roh> stekern: thanks
<lekernel> so this board has two HDMI outputs, one with line driver and one without?
<roh> i was just a bit cautious... i lost multiple graphicscards to bad drivers
<lekernel> drivers?
<roh> some ati, basically all matrox cards
<lekernel> you mean the driver inside the graphics card got damaged?
<roh> line drivers on dvi get worse over time and fail/limit bandwith
<roh> inside the asic
<stekern> actually, I think it has three, 2 with line drivers and one without. But I might remember wrong
<lekernel> stekern, there's one input too
<stekern> I've only tested the output with driver
<roh> it starts slow with 'no 1920x1600 on high refresh anymore' .. and in the end you can drive vga through 20cm of cable but it fails to sync on 120cm
<lekernel> mh, actually the unbuffered one is called "PMODA"... what's that?
<stekern> ah, yeah, that's how it was 2 line in and 2 line out
<stekern> PMODA is digilents extension "standard", there is a hdmi connector in parallell with that
<lekernel> "use of the low-cost TMDS buffer TMDS141 from Texas Instruments to offer additional signal clean-up and ESD protection up to 6 kV"
<lekernel> and TMDS output is only available in banks 0 and 2
<lekernel> with VCCO=3.3V...
<lekernel> on M1 we have plenty of suitable I/Os available in bank 2
<lekernel> unfortunately they're on the opposite side of the BGA, so there can be slight routing difficulties
<lekernel> bank 0 is potentially easier to route, but there's only one I/O left there
<lekernel> hmm... are the 50 ohm series resistors necessary?
<wpwrak> hmm, http://rubidium.dyndns.org/pipermail/fpga-synth/2011-April/001667.html only talks about frequency but not about cable length