That's the way it should be. If you want to subscribe to something, use RSS. That's totally under the control of the recipient. If you unsubscrbe from an RSS feed, there's no way the sender can keep sending to you.
It's easy to follow an RSS feed if you're using Thunderbird; a bit harder if you're a Google slave.
Now, from the people who brought you PCMCIA cards... Remember when you could slot an Ethernet interface or a modem into a PCMCIA slot? Same idea.
Phones should be going in the other direction. No connectors at all. A phone today has about four or five radios in it; do any data transfer over WiFi or Bluetooth or the cellular link. Charging should be inductive, which will happen when one of the three competing wireless charging systems wins. Phones should be waterproof, shockproof, dust-resistant, and close to indestructable.
Energy harvesting technology is real, but you don't get much energy unless there's a transmitter nearby, or a really big one in the area. There are now ICs which slowly accumulate energy in a capacitor, and when they have enough, power up some device and run it for a few milliseconds. Depending on RF levels in the area, you may get some useful output several times a second, or several times a day.
Now if you had something like a microwave flashlight (an low power oscillator and horn antenna) you could wave it around and wake up RFID tags, which could then report back. That could be a viable commercial product, more for industrial and commercial than home use. ("Where's a carton of P/N xxxxx-xxxx parts? Oh, there it is, behind that other box.")
This is not good. This is being done by people from "social", where nothing really has to work. It operates in the "deadly valley" - enough automation to allow the driver to take their hands off the wheel, but not enough automation to handle hard situations. Most of the major auto manufacturers already have that working. Toyota calls it "Lane keeping assist". It's coupled with "smart cruise control", which measures the distance to the car ahead and controls speed and braking. Ford, Mercedes, Volkswagen, and Cadillac have similar systems. Audi already has such as system. But Audi won't let it take full control. "The driver is still responsible", they say. Audi disengages their system if the driver takes their hands off the wheel.
So this is a known technology which none of the major automakers trust enough to give it full control of the vehicle. That should tell you something.
"Cruise is currently taking pre-orders for its first system." Typical. Is there a Kickstarter, too?
These seem similar to the Nao, which is a line of small humanoid robots from France. About the same price point.
What you can do with them depends strongly on the sensors. If the joints are position-controlled only, and you don't have force feedback, locomotion and manipulation will be clunky. There are some simple robot components, such as 6 degree of freedom force sensors for wrists and ankles, which are insanely expensive today, because they're made by hand for research and industrial purposes.
If you have all that sensing, plus three axes of accelerometer and three axes of rate gyro, you should be able to get Boston Dynamics type agility out of the thing. If the DARPA Humanoid Challenge produces some usable open-source software, it should be possible to move that down to toy-sized robots.
This is cool, and certainly buildable, but probably not too useful. It has all the problems of Personal Rapid Transit systems, plus the problems of suspended monorails, plus the problems of maglev.
PRT systems are cool, but to accomodate a lot of people going to different places, you need a lot of stations and track. If a lot of people are going to the same place, bus/railroad car vehicles are more effective. Lots of airports have tracked tram systems with vehicles/trains in the 10-100 passenger range, but none have two or four seaters. It's this scaling problem that's killed PRT systems.
Here's a small-car PRT system that's just about to open in Korea. It's more of an amusement park ride than a transportation system. Note that the guideway is much heavier duty than in the proposed maglev system. This is typical of monorails that get built vs. monorails in pretty pictures. Once you deal with all the real-world problems, like high wind stresses, a truck hitting a support pole, being able to evacuate people from stalled vehicles, and such, the components get bigger. Compare proposed LA monorail from 1950s with actual Chiba monotrail from 1970s
As an earthmoving project, each kilometer of wall is 18M cubic meters. The Panama Canal was about 250M cubic meters of earthmoving. So every 14KM of wall is one Panama Canal. The proposed Arabian Canal near Dubai (to create "valuable waterfront property" accessable by yacht) would require about 1100M cubic meters of earthmoving. So one Arabian Canal is about 60KM of wall.
In terms of speed, one Bagger 288 can move about 250K cubic meters of earth a day. That's 5KM of wall per year. With one such $100 million machine for every 100KM of wall, the project would take 20 years.
It's a big project, but not impossibly big. Just expensively big.
Procedural generation of outdoor scenes has been done for years. Decades, even. Works fine. Most of nature can be simulated with fractals, and basic terrain generators are simple. Speedtree turns out really nice trees and vegetation.
What's really tough is procedural generation of cities. There are programs that build a skyline, but so far, nobody has been able to procedurally generate a convincing city at high-detail level. There are systems that tried, like Introversion and Instant Architecture, but the sameness of the buildings makes for boring cities. It's easy to do this badly, but very hard to do it well.
It's not impossible. You'd need something like The Sims engine or SimCity, which grows cities over time in response to their occupants'' needs. That would be a big win. You could build something like GTA without an army of people constructing the real estate.
But what if my home is out in the middle of the forest, miles from anything resembling internet connectivity... That's a home I want automated more than the home I sleep in the other 5 days a week.
Right. Reporting is from buildings that are mostly unoccupied is really useful. Industrial facilities have used that for decades - unattended pumping stations, power substations, water level gauges, and storage buildings with air conditioning routinely phone home. They usually have very limited bandwidth - pager channels are often used. Usually, they send a message every few minutes with a few numbers and an "I'm fine" message. If there's trouble, they start sending alarm messages. This is the real, existing "Internet of Things", but it's called "M2M" (machine to machine) in the industry. Two-way pager channels, cellular-based pagers, and two way satellite links are used. So your pipeline pumping station can phone home from Outer Nowhere, or you can keep track of what the irrigation pumps on the far side of your farm are doing, or your vacation cabin can check in.
Lucas originally wanted this location for his monument to himself. It's in open space in the Golden Gate National Recreation Area, with unobstructed views of the Golden Gate Bridge, the San Francisco Bay, and Alcatraz. That's one of the great urban waterfront public spaces of the world. So that plan was very unpopular in SF. Another location was proposed, some unused piers a few miles down the waterfront in the tourist waterfront area, but that wasn't dramatic enough for Lucas.
Here's the API reference. It doesn't let you see or do much. I though the Nest was supposed to "learn" your behavior patterns, but if it does, that info isn't exposed in the API. You can look at the temperature and heat/cool/fan status, and maybe change the setpoints. You can tell if someone is home, and when they set the time for when they were coming back.
This isn't an API for the device. It's an API for a Google-hosted service that controls the device. Google is in total control of your home.
Actually, the real profit is in licensing the shovels with a per-scoop fee.
Yes. See Howard Hughes, Sr. (Not the airplane guy; his father, the oil-drilling guy.) He licensed the only good oil-well drill bits for a fraction of the revenue from the well.
As a 450GHz computing element, this is a long way off.
But it might lead to better terahertz radar. Right now, operating in the terahertz range is painfully difficult. It's a strange region where both electronics and optics work, but not easily. This may be a more effective way to work in that range.
I pay my ISP (sonic.net) for Internet access and an IMAP account on an email server. They pay for backbone bandwidth. No advertising revenue is involved.
I pay HostGator a few dollars a month to host my web sites. No advertising revenue is involved. Hosting is amazingly cheap. Basic hosting today is cheaper than remote backup for your smartphone. If I want to put up an image, it goes on Hostgator, not some "sharing" site.
Google's original business model was to charge ISPs for access to their search engine. If I had to pay a few dollars a month for that, it would be fine. If Google dumped ads, "search personalization" and "social", the search engine would be much cheaper to run, and they could lay off most of their sales force.
Before we had "social networks", we had Usenet, which is completely distributed, redundant, and ad-free. Still around, too.
Maps, though. Those are expensive. Surprisingly, most of the map systems aren't ad-heavy.
This is a nice little trick. This has the potential to extend shared consistent memory multiprocessor designs to far larger numbers of processors. Whether this is a performance win remains to be seen. Good idea, though. Note that the prototype chip is just a feasibility test; they used an off the shelf Power CPU design, added their interconnect network, and send the job off to a fab. A production chip would have optimizations this does not.
We known only two general purpose multiprocessor architectures that are broadly useful - shared consistent memory multiprocessors, and clusters of machines with no shared memory. Dozens of other schemes have been tried - SIMD machines (the Connection Machine), non-shared memory with DMA to a bigger memory (the Cell), message passing to adjacent machines in N dimensions (Hypercube), message passing over an on-chip network (several examples), cross-CPU DMA access (Infiniband) and shared memory without cache consistency (Intel experimental). In all cases, the hardware worked, the programming was a problem from hell, and the concept was dropped. The Cell in the PS3 is the only high-volume product with an exotic multiprocessor architecture, and that was such a pain that the PS4 dropped it for a more conventional architecture.
It's a good thing that some of those weapons were brought to the prototype stage, but not to production. Today, there's a strong tendency to have only one program underway for major aircraft, leading to production of marginal aircraft like the F-35.
There are many smaller weapons, such as the XM8 assault rifle, which made it to prototype but were then cancelled. Guided ammo for small arms has been demonstrated, but it's still some ways from being miitarily useful.
Laser weapons are in the same state - there are working demos, but they're not worth the trouble yet. Diode laser powered weapons are now up to 10KW (big array of 10W or so diodes), and can shoot down small rockets and artillery shells in demos. Current thinking is that, at 50KW-100KW, they'll be militarily useful.
Blackberry - only vendor serious about security
on
BlackBerry Back In Profit
·
· Score: 4, Insightful
Blackberry may make a comeback as the "big business smartphone". All the other smartphones are slaves to Apple or Google or a carrier. Blackberry phones are slaves to the enterprise Blackberry server, and Blackberry itself doesn't see phone traffic. Blackberry is the only major vendor serious about security and encryption. Everybody else is into advertising revenue.
Unless there's some well organized exchange, as for stocks and bonds, the U.S. Marshals Service sells everything by auction. No Bitcoin exchange is solid enough for a transaction of this size.
Bank of America's online bill-paying system is quite good. They handle the interfacing with other sites, and most larger companies are interconnected with them. You can view incoming bills on the BofA site and pay them there.
This doesn't let other companies debit your account. You have to initiate transfers while logged into BofA. So this is much safer than letting companies initiate transfers to them from their end.
We pretty much have everything in place to replace first class mail. 98% of the paper volume in my mailbox is in the recycling bin immediately after receipt.
Right. I've been there, but back in 2002, before they got real money. The people giving tours are obsessed with the buildings. The sheds are, indeed, sheds - temporary buildings built during WWII. Even the manor house isn't architecturally significant. It's only a Grade II listed building, along with about 300,000 other Grade II listed buildings in the UK.
Nobody seems to be interested in the room at the Institute for Advanced Studies in Princeton where von Neumann had the EDVAC built. It's a storeroom now. The EDVAC was a real general purpose computer, with the "von Neumann architecture" all later computers followed.
"Colossus" and the bombes were special-purpose machines for crypto key testing, like the ASICs used for Bitcoin mining. They really belong to the separate history of digital special-purpose machines, such as SIGSALY (digital voice encryption), Reservisor (travel reservations), American Totalizators's racetrack hardware, Teleregister's stock exchange and inventory control hardware, and Western Union's message switching systems. All that stuff is obsolete and forgotten.
Uh, I thought this was the descendant of Burroughs B5000? You know, the computer that Alan Kay tells everyone to take a look at to understand how silly today's architectures look in comparison.
That's the other Unisys line; they have an A-series (from Burroughs) and a B-series (from UNIVAC).
I used a B5500, at UC Santa Cruz, in a summer course on computer architecture in 1975, taught by one of its designers. Burroughs donated the obsolete machine, and we stepped it through instructions from its maintenance panel, watching the stack hardware work. We were also taken up to Xerox PARC to meet Alan Kay and see the original prototype Alto machines, years before Steve Jobs did. (They were really Data General Nova machines inside, with different microcode.)
Burroughs machines never matched C well. It's an assumption of C that the memory is one big numbered address space. That's not how the Burroughs machines work. Memory, as seen by user programs, is a tree, like a file system. A local variable has an address something like Program7.Function23.Variable3(numeric). As with a file system, you don't know the underlying hardware address.
The basic type on those machines is a 48-bit number. Integers and floating point numbers have the same format. The word has a sign, an exponent sign, an exponent, and a mantissa. The binary point is at the low end. So integers up to 2^32 and floats have the same representation. The hardware will convert a result to float when it overflows an integer. This, again, is something C isn't ready for.
That's over half a century of the UNIVAC 36-bit architecture, going back to the UNIVAC 1107. The operating system in use today, originally EXEC 8, later OS 1100, later OS 2200, first ran on the UNIVAC 1108 in 1966.
Some programs from the 1970s will still run today. Some from that era are still being maintained and distributed.
Slashdot Mirror
That's the way it should be. If you want to subscribe to something, use RSS. That's totally under the control of the recipient. If you unsubscrbe from an RSS feed, there's no way the sender can keep sending to you.
It's easy to follow an RSS feed if you're using Thunderbird; a bit harder if you're a Google slave.
Now, from the people who brought you PCMCIA cards... Remember when you could slot an Ethernet interface or a modem into a PCMCIA slot? Same idea.
Phones should be going in the other direction. No connectors at all. A phone today has about four or five radios in it; do any data transfer over WiFi or Bluetooth or the cellular link. Charging should be inductive, which will happen when one of the three competing wireless charging systems wins. Phones should be waterproof, shockproof, dust-resistant, and close to indestructable.
Energy harvesting technology is real, but you don't get much energy unless there's a transmitter nearby, or a really big one in the area. There are now ICs which slowly accumulate energy in a capacitor, and when they have enough, power up some device and run it for a few milliseconds. Depending on RF levels in the area, you may get some useful output several times a second, or several times a day.
Now if you had something like a microwave flashlight (an low power oscillator and horn antenna) you could wave it around and wake up RFID tags, which could then report back. That could be a viable commercial product, more for industrial and commercial than home use. ("Where's a carton of P/N xxxxx-xxxx parts? Oh, there it is, behind that other box.")
This is not good. This is being done by people from "social", where nothing really has to work. It operates in the "deadly valley" - enough automation to allow the driver to take their hands off the wheel, but not enough automation to handle hard situations. Most of the major auto manufacturers already have that working. Toyota calls it "Lane keeping assist". It's coupled with "smart cruise control", which measures the distance to the car ahead and controls speed and braking. Ford, Mercedes, Volkswagen, and Cadillac have similar systems. Audi already has such as system. But Audi won't let it take full control. "The driver is still responsible", they say. Audi disengages their system if the driver takes their hands off the wheel.
So this is a known technology which none of the major automakers trust enough to give it full control of the vehicle. That should tell you something.
"Cruise is currently taking pre-orders for its first system." Typical. Is there a Kickstarter, too?
These seem similar to the Nao, which is a line of small humanoid robots from France. About the same price point.
What you can do with them depends strongly on the sensors. If the joints are position-controlled only, and you don't have force feedback, locomotion and manipulation will be clunky. There are some simple robot components, such as 6 degree of freedom force sensors for wrists and ankles, which are insanely expensive today, because they're made by hand for research and industrial purposes.
If you have all that sensing, plus three axes of accelerometer and three axes of rate gyro, you should be able to get Boston Dynamics type agility out of the thing. If the DARPA Humanoid Challenge produces some usable open-source software, it should be possible to move that down to toy-sized robots.
This is cool, and certainly buildable, but probably not too useful. It has all the problems of Personal Rapid Transit systems, plus the problems of suspended monorails, plus the problems of maglev.
PRT systems are cool, but to accomodate a lot of people going to different places, you need a lot of stations and track. If a lot of people are going to the same place, bus/railroad car vehicles are more effective. Lots of airports have tracked tram systems with vehicles/trains in the 10-100 passenger range, but none have two or four seaters. It's this scaling problem that's killed PRT systems.
Here's a small-car PRT system that's just about to open in Korea. It's more of an amusement park ride than a transportation system. Note that the guideway is much heavier duty than in the proposed maglev system. This is typical of monorails that get built vs. monorails in pretty pictures. Once you deal with all the real-world problems, like high wind stresses, a truck hitting a support pole, being able to evacuate people from stalled vehicles, and such, the components get bigger. Compare proposed LA monorail from 1950s with actual Chiba monotrail from 1970s
As an earthmoving project, each kilometer of wall is 18M cubic meters. The Panama Canal was about 250M cubic meters of earthmoving. So every 14KM of wall is one Panama Canal. The proposed Arabian Canal near Dubai (to create "valuable waterfront property" accessable by yacht) would require about 1100M cubic meters of earthmoving. So one Arabian Canal is about 60KM of wall.
In terms of speed, one Bagger 288 can move about 250K cubic meters of earth a day. That's 5KM of wall per year. With one such $100 million machine for every 100KM of wall, the project would take 20 years.
It's a big project, but not impossibly big. Just expensively big.
Procedural generation of outdoor scenes has been done for years. Decades, even. Works fine. Most of nature can be simulated with fractals, and basic terrain generators are simple. Speedtree turns out really nice trees and vegetation.
What's really tough is procedural generation of cities. There are programs that build a skyline, but so far, nobody has been able to procedurally generate a convincing city at high-detail level. There are systems that tried, like Introversion and Instant Architecture, but the sameness of the buildings makes for boring cities. It's easy to do this badly, but very hard to do it well.
It's not impossible. You'd need something like The Sims engine or SimCity, which grows cities over time in response to their occupants'' needs. That would be a big win. You could build something like GTA without an army of people constructing the real estate.
But what if my home is out in the middle of the forest, miles from anything resembling internet connectivity... That's a home I want automated more than the home I sleep in the other 5 days a week.
Right. Reporting is from buildings that are mostly unoccupied is really useful. Industrial facilities have used that for decades - unattended pumping stations, power substations, water level gauges, and storage buildings with air conditioning routinely phone home. They usually have very limited bandwidth - pager channels are often used. Usually, they send a message every few minutes with a few numbers and an "I'm fine" message. If there's trouble, they start sending alarm messages. This is the real, existing "Internet of Things", but it's called "M2M" (machine to machine) in the industry. Two-way pager channels, cellular-based pagers, and two way satellite links are used. So your pipeline pumping station can phone home from Outer Nowhere, or you can keep track of what the irrigation pumps on the far side of your farm are doing, or your vacation cabin can check in.
Lucas originally wanted this location for his monument to himself. It's in open space in the Golden Gate National Recreation Area, with unobstructed views of the Golden Gate Bridge, the San Francisco Bay, and Alcatraz. That's one of the great urban waterfront public spaces of the world. So that plan was very unpopular in SF. Another location was proposed, some unused piers a few miles down the waterfront in the tourist waterfront area, but that wasn't dramatic enough for Lucas.
Here's the API reference. It doesn't let you see or do much. I though the Nest was supposed to "learn" your behavior patterns, but if it does, that info isn't exposed in the API. You can look at the temperature and heat/cool/fan status, and maybe change the setpoints. You can tell if someone is home, and when they set the time for when they were coming back.
This isn't an API for the device. It's an API for a Google-hosted service that controls the device. Google is in total control of your home.
Actually, the real profit is in licensing the shovels with a per-scoop fee.
Yes. See Howard Hughes, Sr. (Not the airplane guy; his father, the oil-drilling guy.) He licensed the only good oil-well drill bits for a fraction of the revenue from the well.
As a 450GHz computing element, this is a long way off. But it might lead to better terahertz radar. Right now, operating in the terahertz range is painfully difficult. It's a strange region where both electronics and optics work, but not easily. This may be a more effective way to work in that range.
email
I pay my ISP (sonic.net) for Internet access and an IMAP account on an email server. They pay for backbone bandwidth. No advertising revenue is involved.
I pay HostGator a few dollars a month to host my web sites. No advertising revenue is involved. Hosting is amazingly cheap. Basic hosting today is cheaper than remote backup for your smartphone. If I want to put up an image, it goes on Hostgator, not some "sharing" site.
Google's original business model was to charge ISPs for access to their search engine. If I had to pay a few dollars a month for that, it would be fine. If Google dumped ads, "search personalization" and "social", the search engine would be much cheaper to run, and they could lay off most of their sales force.
Before we had "social networks", we had Usenet, which is completely distributed, redundant, and ad-free. Still around, too.
Maps, though. Those are expensive. Surprisingly, most of the map systems aren't ad-heavy.
This is a nice little trick. This has the potential to extend shared consistent memory multiprocessor designs to far larger numbers of processors. Whether this is a performance win remains to be seen. Good idea, though. Note that the prototype chip is just a feasibility test; they used an off the shelf Power CPU design, added their interconnect network, and send the job off to a fab. A production chip would have optimizations this does not.
We known only two general purpose multiprocessor architectures that are broadly useful - shared consistent memory multiprocessors, and clusters of machines with no shared memory. Dozens of other schemes have been tried - SIMD machines (the Connection Machine), non-shared memory with DMA to a bigger memory (the Cell), message passing to adjacent machines in N dimensions (Hypercube), message passing over an on-chip network (several examples), cross-CPU DMA access (Infiniband) and shared memory without cache consistency (Intel experimental). In all cases, the hardware worked, the programming was a problem from hell, and the concept was dropped. The Cell in the PS3 is the only high-volume product with an exotic multiprocessor architecture, and that was such a pain that the PS4 dropped it for a more conventional architecture.
It's a good thing that some of those weapons were brought to the prototype stage, but not to production. Today, there's a strong tendency to have only one program underway for major aircraft, leading to production of marginal aircraft like the F-35.
There are many smaller weapons, such as the XM8 assault rifle, which made it to prototype but were then cancelled. Guided ammo for small arms has been demonstrated, but it's still some ways from being miitarily useful.
Laser weapons are in the same state - there are working demos, but they're not worth the trouble yet. Diode laser powered weapons are now up to 10KW (big array of 10W or so diodes), and can shoot down small rockets and artillery shells in demos. Current thinking is that, at 50KW-100KW, they'll be militarily useful.
Blackberry may make a comeback as the "big business smartphone". All the other smartphones are slaves to Apple or Google or a carrier. Blackberry phones are slaves to the enterprise Blackberry server, and Blackberry itself doesn't see phone traffic. Blackberry is the only major vendor serious about security and encryption. Everybody else is into advertising revenue.
Why don't they just exchange them for USD?
Unless there's some well organized exchange, as for stocks and bonds, the U.S. Marshals Service sells everything by auction. No Bitcoin exchange is solid enough for a transaction of this size.
Communications of the Association for Computing Machinery was first published in 1957.
A higher gasoline tax beats the proposed alternative, monitoring all cars for distance driven.
Bank of America's online bill-paying system is quite good. They handle the interfacing with other sites, and most larger companies are interconnected with them. You can view incoming bills on the BofA site and pay them there.
This doesn't let other companies debit your account. You have to initiate transfers while logged into BofA. So this is much safer than letting companies initiate transfers to them from their end.
We pretty much have everything in place to replace first class mail. 98% of the paper volume in my mailbox is in the recycling bin immediately after receipt.
All of which you can do with FedEx, UPS, or the USPS's express flat service. It costs more, but how many times a year do you use that service?
Other than for bills, first class mail is dead. For bills, it's dying.
Right. I've been there, but back in 2002, before they got real money. The people giving tours are obsessed with the buildings. The sheds are, indeed, sheds - temporary buildings built during WWII. Even the manor house isn't architecturally significant. It's only a Grade II listed building, along with about 300,000 other Grade II listed buildings in the UK.
Nobody seems to be interested in the room at the Institute for Advanced Studies in Princeton where von Neumann had the EDVAC built. It's a storeroom now. The EDVAC was a real general purpose computer, with the "von Neumann architecture" all later computers followed.
"Colossus" and the bombes were special-purpose machines for crypto key testing, like the ASICs used for Bitcoin mining. They really belong to the separate history of digital special-purpose machines, such as SIGSALY (digital voice encryption), Reservisor (travel reservations), American Totalizators's racetrack hardware, Teleregister's stock exchange and inventory control hardware, and Western Union's message switching systems. All that stuff is obsolete and forgotten.
Uh, I thought this was the descendant of Burroughs B5000? You know, the computer that Alan Kay tells everyone to take a look at to understand how silly today's architectures look in comparison.
That's the other Unisys line; they have an A-series (from Burroughs) and a B-series (from UNIVAC).
I used a B5500, at UC Santa Cruz, in a summer course on computer architecture in 1975, taught by one of its designers. Burroughs donated the obsolete machine, and we stepped it through instructions from its maintenance panel, watching the stack hardware work. We were also taken up to Xerox PARC to meet Alan Kay and see the original prototype Alto machines, years before Steve Jobs did. (They were really Data General Nova machines inside, with different microcode.)
Burroughs machines never matched C well. It's an assumption of C that the memory is one big numbered address space. That's not how the Burroughs machines work. Memory, as seen by user programs, is a tree, like a file system. A local variable has an address something like Program7.Function23.Variable3(numeric). As with a file system, you don't know the underlying hardware address.
The basic type on those machines is a 48-bit number. Integers and floating point numbers have the same format. The word has a sign, an exponent sign, an exponent, and a mantissa. The binary point is at the low end. So integers up to 2^32 and floats have the same representation. The hardware will convert a result to float when it overflows an integer. This, again, is something C isn't ready for.
That's over half a century of the UNIVAC 36-bit architecture, going back to the UNIVAC 1107. The operating system in use today, originally EXEC 8, later OS 1100, later OS 2200, first ran on the UNIVAC 1108 in 1966.
Some programs from the 1970s will still run today. Some from that era are still being maintained and distributed.