The biggest problem with CIGS as a production material is probably that it can't "piggyback" on the industry built up for the computer industry. I know that sounds strange, since its lack of reliance on that source of material is also its advantage, but tools to work with CIGS have to be developed more or less from scratch. That's expensive, and the reason that these initial investments are important. The process must be bootstrapped.
I would argue, however, that this is actually partly an advantage. The semiconductor industry is based on batch processes for the most part. That usually means slow. Miasole's factory uses reel-to-reel manufacturing, and will be able to exploit the high volume equipment designed for this type of manufacturing.
I've seen the Miasole production facility and had a chat with the CEO and one of the engineers at the end of the summer. There're a few interesting things that TFA doesn't mention. First, Miasole claims the low $25M price tag for a 200MW factory because they build all of their equipment from scratch. When I was on the floor, they were building a single 25MW line which they turned on for testing last month. That cost them a grand total of $4M (in parts) to build. E.g. they've already done one, so the pricing is reasonably accurate. Subsequent lines will be cheaper. This will give them a huge cost advantage over other similar companies.
Secondly, their production process is cheaper not only because material costs are lower, but also because they use a "reel-to-reel" process in which the semiconductor material is deposited on a sheet of steel which unrolls into the line, and then rolls back up on a reel on the other side. The steel sheets can then be cut and woven into a vinyl enclosure which can be rolled out on your roof like regular roofing shingles. Cool stuff. (They're probably going to attack industrial markets first though...)
Third, the management team comes from the disk drive industry, and built the Seagate facility that is responsible for ~30% of the world's hard drives (could have the percentage slightly wrong, but is in the ballpark). Hard drives use a similar thin film deposition process, and they have built several other manufacturing systems based on thin film processes. This is why the are able to get such a low cost on their equipment: they have the contacts and expertise to build from scratch.
For the record, I have not talked with their competitors, so I don't know the whole story, but Miasole seems very well positioned, and their facility is certainly real.
Actually, there are several other things that Silicon Valley has. First, it has by far the most of the VC money: 34% of the VC money in the US is invested in Silicon valley, and around 90% of the world's VC is invested in the US. Secondly, as TFA points out, silicon valley is near SF, a nice place to live. Third, it is close to Stanford and Berkeley, two of the worlds formost engineering schools. Fourth, it has a highly trained workforce which understands both tech and entrepreurship. Most importantly, silicon valley has a very mobile workforce, which moves around companies, starts companies and takes its institutional knowledge with it. This "cultural" characteristic is probably the single most important differentiating factor between Silicon Valley and other locations.
AnnaLee Saxenian has a great book on this topic called "Regional Advantage." It goes into these issues in great detail.
Also, VOIP apps would probably just eventually adapt by providing their own QoS using overlay networks (for example OverQoS) if QoS became a real issue.
Some folks at Santa Cruz have been working on this type of thing for a while now, except that the techniques the guys at UCSC are using are adaptive: as the workload changes, so does the algorithm. I doubt these guys could do better with an offline algorithm.
Another related item is the ARC algorithm from IBM, which is an adaptive cache for block buffers.
Well, this patent builds on another patent owned by the same authors: US Patent 4,302,775. This patent describes a technique for "Two Dimensional Discrete Cosine Transformation", the technique which JPEG uses to compress images:
Another object of the present invention is to provide a digital video compression system which effectively implements a two-dimensional discrete cosine transform of blocks of the picture.
It seems like the press release should have pointed to this patent, rather than the one cited, but assuming that their lawyers got it right, there may be more of a case here.
IANAL, but I doubt that you need to mention all the the potential uses for the techniques you describe. Just because a technique only uses a subset of the ideas in your patent doesn't necessarily mean (to me) that the idea is not covered by your patent.
Clearly, their behavior is shady, but I suspect that they actually have a case (at least based on the technology). Can someone clarify this?
Well, most of the storage vendors are moving towards using cheap hard disks as archival storage (for example Network Appliance's NearStore), and I expect we will soon begin to see this kind of thing for smaller scale operations, so I doubt there is much market for this except for consumers. With the new 27GB DVDs due out sometime soon, I doubt there's much market there either.
If they could come up with a 350MB or 3.5TB removable drive at a reasonable price, they might have a viable idea. But as it stands, with their reputation in tatters, I doubt this will go very far.
When processing on a transaction basis, the transactions often need to posted in the exact order they were recieved, which means they must be taken serially.
Indeed this is true, but even in this case most applications can achieve parallelism via pipelining. In other words, serial events can often be handled in multiple steps, in the same way that a processor pipelines. This is common for query processing in distributed relational databases.
It still depends on the type of application though. Some applications are also difficult to pipeline.
Actually, its not so much the amount of data that needs to be analyzed, input and output, but the interdependence of data.
For example, in the SETI project, there is a hugh amount of data which needs to be analyzed, but one chunk of work has no dependence on another, so the work can be spread to thousands of workstations all over the world. This is the hallmark of an "embarssingly parallel" problem.
Essentially because people are economically disadvantage when they don't have equal access. The point is not so people with disabilities can feel normal. The point is so that people can be functional and independent, despite their disability.
The job wears on them all, day after day, so much so that the obscenity division has recently set up in-house counseling for them to talk about what they're seeing and how it is affecting them.
"This stuff isn't the easiest to deal with," Nguyen said recently while at his computer. "But I think we're going after the bad guys and we're making a difference, and that's what makes it worthwhile."
I can hear it now: "Three women... but the worst part was... (sob) I... liked... it."
How ever do they find people to do work like this?
Slashdot Mirror
The biggest problem with CIGS as a production material is probably that it can't "piggyback" on the industry built up for the computer industry. I know that sounds strange, since its lack of reliance on that source of material is also its advantage, but tools to work with CIGS have to be developed more or less from scratch. That's expensive, and the reason that these initial investments are important. The process must be bootstrapped. I would argue, however, that this is actually partly an advantage. The semiconductor industry is based on batch processes for the most part. That usually means slow. Miasole's factory uses reel-to-reel manufacturing, and will be able to exploit the high volume equipment designed for this type of manufacturing.
I've seen the Miasole production facility and had a chat with the CEO and one of the engineers at the end of the summer. There're a few interesting things that TFA doesn't mention. First, Miasole claims the low $25M price tag for a 200MW factory because they build all of their equipment from scratch. When I was on the floor, they were building a single 25MW line which they turned on for testing last month. That cost them a grand total of $4M (in parts) to build. E.g. they've already done one, so the pricing is reasonably accurate. Subsequent lines will be cheaper. This will give them a huge cost advantage over other similar companies.
Secondly, their production process is cheaper not only because material costs are lower, but also because they use a "reel-to-reel" process in which the semiconductor material is deposited on a sheet of steel which unrolls into the line, and then rolls back up on a reel on the other side. The steel sheets can then be cut and woven into a vinyl enclosure which can be rolled out on your roof like regular roofing shingles. Cool stuff. (They're probably going to attack industrial markets first though...)
Third, the management team comes from the disk drive industry, and built the Seagate facility that is responsible for ~30% of the world's hard drives (could have the percentage slightly wrong, but is in the ballpark). Hard drives use a similar thin film deposition process, and they have built several other manufacturing systems based on thin film processes. This is why the are able to get such a low cost on their equipment: they have the contacts and expertise to build from scratch.
For the record, I have not talked with their competitors, so I don't know the whole story, but Miasole seems very well positioned, and their facility is certainly real.
Actually this is the standard measure of factory capacity for the solar industry, and is explained in TFA.
Actually, there are several other things that Silicon Valley has. First, it has by far the most of the VC money: 34% of the VC money in the US is invested in Silicon valley, and around 90% of the world's VC is invested in the US. Secondly, as TFA points out, silicon valley is near SF, a nice place to live. Third, it is close to Stanford and Berkeley, two of the worlds formost engineering schools. Fourth, it has a highly trained workforce which understands both tech and entrepreurship. Most importantly, silicon valley has a very mobile workforce, which moves around companies, starts companies and takes its institutional knowledge with it. This "cultural" characteristic is probably the single most important differentiating factor between Silicon Valley and other locations.
AnnaLee Saxenian has a great book on this topic called "Regional Advantage." It goes into these issues in great detail.
I currently have a professor who bans laptops for essentially the same reason, and in my opinion it leads to better discussion in class.
http://news-service.stanford.edu/news/2005/june15/ jobs-061505
Also, VOIP apps would probably just eventually adapt by providing their own QoS using overlay networks (for example OverQoS) if QoS became a real issue.
Another related item is the ARC algorithm from IBM, which is an adaptive cache for block buffers.
IANAL, but I doubt that you need to mention all the the potential uses for the techniques you describe. Just because a technique only uses a subset of the ideas in your patent doesn't necessarily mean (to me) that the idea is not covered by your patent. Clearly, their behavior is shady, but I suspect that they actually have a case (at least based on the technology). Can someone clarify this?
If they could come up with a 350MB or 3.5TB removable drive at a reasonable price, they might have a viable idea. But as it stands, with their reputation in tatters, I doubt this will go very far.
It still depends on the type of application though. Some applications are also difficult to pipeline.
For example, in the SETI project, there is a hugh amount of data which needs to be analyzed, but one chunk of work has no dependence on another, so the work can be spread to thousands of workstations all over the world. This is the hallmark of an "embarssingly parallel" problem.
Essentially because people are economically disadvantage when they don't have equal access. The point is not so people with disabilities can feel normal. The point is so that people can be functional and independent, despite their disability.
How ever do they find people to do work like this?
See the Berkeley TIER Project and the Delay Tolerant Networking Research Group"
There is a java client called jCIFS that is part of the Samba project.