Quote:
In bulk we can get it cheaper (we are actively seeking to buy this components)...
That is good to know!
Quote:
...but for a single unit price easily available on the web, $30 is pretty much how much these risers cost.
For a single board and a single project it is not a significant outlay at all :)
True enough. I had not actually shopped around for the risers yet.. When you cat get much more complicated cards with integrated circuit chips and everything for $5 bucks, like a
Realtek 8139D NIC for instance, it just "felt" expensive. But then it is a "supply and demand" issue, I suppose. A case of specialty-item-itis resulting in an enlarged priceus-maximus. Take two bulk orders and call me in the morning....
Quote:
It's an option but we are still evaluating chipsets. Consider the power consumption and featureset of the XGI Volari; MPEG2 decoder, what we might call "DirectX 8" class graphics. That is basically the mid-range of what you need to run Windows XP from a graphics adapter point of view.
For a very low power client, we can obtain chipsets and solutions for graphics which cost a third of the usual price of these class chipsets, with little/no 3D or video decoding offload, which draw 1/10th of the power (think of nVidia or ATI mobile device chipsets - not Mobility Radeon, but mobile phone!). This would suit portable or battery-operated devices much better.
...snip...
Quote:
Since your project is not truly graphics-oriented it shouldn't really matter which card you use from the standpoint of getting a display and showing the data and mesh networking. The Volari is useful to gain HWMC MPEG2 decoding (I don't think they will be watching DVD movies while they rescue people from earthquakes :) and the 3D and HDTV (component video) support.
Very very true! My project would not need anything, graphically speaking, more than what would be necessary to support basic web browsing. I won't even care about streaming video, though I suppose it would be nice if it could support that. But advanced 3D graphics??? MPEG-2 decoding (or encoding)??? Nope. Not needed. At a stretch, perhaps, someone will want to somehow pipe back video shot "on the ground". But even that would be considered superfluous as far as project goals are concerned. (Maybe we'll give them
HASCII-CAM, lol!)
hasciicam
Yeah, it's probably what you think it is... Jaromil develops under the RASTASOFT moniker. He makes the dyne:bolic distro. This is also where I first heard of LFS (Linux From Scratch), as dyne:bolic is (or was) developed from it. And whatever your feelings about ganga, genuine rastafarians, judging from those I have met, tend to exhibit outstanding character as people IMHO. Then again, luvdownbabylon should tell you something of my beliefs, even though I don't consider myself a rasta proper.
Now, streaming and encoding voice; that
is a priority. I have every intention of giving the kiosk units VoIP capabilities.
The
one and only reason I would really care about working with the XGI Volari cards, from a project standpoint, is if the EFIKA were to include the same (or compatible) chipset on it's graphics-enabled incarnation. An EFIKA with built-in graphics would, I presume, bring in a lower BoM than one with a riser and a video card. Power requirements would be lower, too, FWIW.
Well, in light of your comments about the PCI riser supply deal, I suppose a secondary reason to develop on the XGI Volari boards would be if Genesi were to have some special pricing worked out that made them a good choice for multi-head kiosks. The point, for me, is to develop with the end product in mind. That's all. Why screw around with a Radeon if I already know it won't be in the end product? As for now, though, that final design is still an open question. So the Radeon suffices.
Quote:
The next Efika should use the next version of the MPC5200B, and will aim add features to the board without significantly increasing cost or removing anything people know and love. I hope the PCI slot stays, at least, it's quite useful.
Oh me too!! The PCI slot is a real blessing to have, IMHO. However my own project's needs would be better served by a miniPCI connection. But for the EFIKA development community as a whole, I think the PCI slot is a real winning feature. And my vote (are we voting? ;) would always lean toward what is better for the community rather than what would serve the narrow interests of my own project. We succeed or fail together.
Speaking of the next EFIKA revision, I believe I read that WiFi is being considered? That could be especially nice. I had originally intended to use a PCI card that gave me two miniPCI connections so that I could include both WiMAX and WiFi radio modules on the MESH units. WiMAX for mesh back-haul, and WiFi as a traditional local AP service allowing WiFi clients (like laptops) access to the net. Including WiFi on a next-gen EFIKA would certainly simplify my project and further reduce the BoM. But, of course, you will design with the greater EFIKA community needs in mind. So I suppose we'll all just have to be patient and see what pet features make it and which don't. (Throw in my vote for additional USB headers!)
Oh and hey!!! Thanks, Matt, for following my project postings. It is good to know you all stay so closely involved with the progress of our individual projects. Maybe it's all mushy gushy but it feels good to know people at Genesi are paying attention! I get a good sense of community here, and I appreciate that. This is my first project, and it is very reassuring.
Peace!
Tim
luvdownbabylon