Since this is the recent open thread...
Mostly for wj, who purports to be an eventual user of what is currently a piece of toy software for dewarping images. After a small frenzy of coding today, here's a very simple-minded cut at color. I chose this image to see if it preserved the red-eye effect in the right eye. (After looking at the original Polaroid print under 5x magnification, this is surprisingly good.) Among the things on my mind as I kept cutting corners were: (a) how many serial color-space and gamma conversions am I ignoring here, and (b) how much information am I losing by forcing intermediate values back to eight-bit integers? Still, I'm not unhappy with the results. http://www.mcain6925.com/obsidian/dewarp/obsidian09.jpg
2025-08-18 12:17:13
Re the link in nous's 3:12...
I'm on the author's side, mostly. So I'll get my initial childish response out of the way: if you're going to argue numbers, for pity's sake format the numbers so they're legible. My normal response when given a table that I have to copy-and-paste into a different piece of software to read conveniently is to just stop there.
Given the title of the linked piece -- "The Humanities as Service Departments: Facing the Budget Logic" -- the author never got to the point I was expecting. What I was looking for was a reconciliation of two facts. On the one side, the author's table showing that A&H generated large amounts of tuition revenue at little expense. On the other, the administration's assertion that A&H departments had relatively few students in their degree programs. He put the conclusion right there in the title, but apparently couldn't bring himself to say it.
It seems to me that A&H faculty should (a) have recognized that they were turning into service departments and (b) have been proactive on the question of how to be better service departments (while retaining their historical roles). The goal ought to be that when someone proposes A&H cuts, the non-A&H faculty scream.
Long ago when I was a TA at the University of Texas, the state legislature proposed what was basically doing away with us and requiring full-time faculty to do the work. I went down to the Capitol the day they had public hearings. Two faculty members killed the bill. First, the head of the math department testified that with his current staff, the dept would have to drop the services they were providing to the engineering school: calculus, linear algebra, and differential equations classes. Second, the head of the engineering college testified that he would pretty much have to shut down if that happened because his faculty would go elsewhere rather than teach the math classes. Worth noting that the math department already segregated students. Both linear algebra and differential equations were taught in two versions, one for engineering students and one for people outside of engineering.
2025-08-17 11:05:03
There's a thing in the education literature called "the Colorado Paradox". We are quite mediocre at getting resident kids through K-12, and into and through college. But we have the 2nd or 3rd most educated workforce in the country. Mostly it's an accident of geography and history, and is probably not reproducible.
2025-08-15 19:55:17
Re TonyP's comment, and a general thought about Alaskan resources... Those would be oil, natural gas, and coal for the most part. Russia already has lots of those.
2025-08-15 19:41:15
Spent some time off and on this week adding a first cut illumination correction to the toy software for processing images of documents.
Original snapshot taken with a handheld iPad here.
Result of correcting various geometry impairments -- curled pages, perspective, orientation -- here.
First cut at doing some illumination correction here. The approach I've taken seems to do well at correcting problems caused purely by the light source being off to the side (darker sections where page curl has the surface normal vector pointing away from the light, brighter where the normal vector is pointing towards the light). The shadow in the lower right corner is from the iPad and will require a different approach to identify and correct. Or more likely, when I get closer to production I'll arrange things so the camera doesn't cast shadows.
Since this is the recent open thread...
Mostly for wj, who purports to be an eventual user of what is currently a piece of toy software for dewarping images. After a small frenzy of coding today, here's a very simple-minded cut at color. I chose this image to see if it preserved the red-eye effect in the right eye. (After looking at the original Polaroid print under 5x magnification, this is surprisingly good.) Among the things on my mind as I kept cutting corners were: (a) how many serial color-space and gamma conversions am I ignoring here, and (b) how much information am I losing by forcing intermediate values back to eight-bit integers? Still, I'm not unhappy with the results.
http://www.mcain6925.com/obsidian/dewarp/obsidian09.jpg
Re the link in nous's 3:12...
I'm on the author's side, mostly. So I'll get my initial childish response out of the way: if you're going to argue numbers, for pity's sake format the numbers so they're legible. My normal response when given a table that I have to copy-and-paste into a different piece of software to read conveniently is to just stop there.
Given the title of the linked piece -- "The Humanities as Service Departments: Facing the Budget Logic" -- the author never got to the point I was expecting. What I was looking for was a reconciliation of two facts. On the one side, the author's table showing that A&H generated large amounts of tuition revenue at little expense. On the other, the administration's assertion that A&H departments had relatively few students in their degree programs. He put the conclusion right there in the title, but apparently couldn't bring himself to say it.
It seems to me that A&H faculty should (a) have recognized that they were turning into service departments and (b) have been proactive on the question of how to be better service departments (while retaining their historical roles). The goal ought to be that when someone proposes A&H cuts, the non-A&H faculty scream.
Long ago when I was a TA at the University of Texas, the state legislature proposed what was basically doing away with us and requiring full-time faculty to do the work. I went down to the Capitol the day they had public hearings. Two faculty members killed the bill. First, the head of the math department testified that with his current staff, the dept would have to drop the services they were providing to the engineering school: calculus, linear algebra, and differential equations classes. Second, the head of the engineering college testified that he would pretty much have to shut down if that happened because his faculty would go elsewhere rather than teach the math classes. Worth noting that the math department already segregated students. Both linear algebra and differential equations were taught in two versions, one for engineering students and one for people outside of engineering.
There's a thing in the education literature called "the Colorado Paradox". We are quite mediocre at getting resident kids through K-12, and into and through college. But we have the 2nd or 3rd most educated workforce in the country. Mostly it's an accident of geography and history, and is probably not reproducible.
Re TonyP's comment, and a general thought about Alaskan resources... Those would be oil, natural gas, and coal for the most part. Russia already has lots of those.
Spent some time off and on this week adding a first cut illumination correction to the toy software for processing images of documents.
Original snapshot taken with a handheld iPad here.
Result of correcting various geometry impairments -- curled pages, perspective, orientation -- here.
First cut at doing some illumination correction here. The approach I've taken seems to do well at correcting problems caused purely by the light source being off to the side (darker sections where page curl has the surface normal vector pointing away from the light, brighter where the normal vector is pointing towards the light). The shadow in the lower right corner is from the iPad and will require a different approach to identify and correct. Or more likely, when I get closer to production I'll arrange things so the camera doesn't cast shadows.