4 ms·

Does anyone have any history on what motivated the industry to commoditize PC hardware? The current trend is in the opposite direction, with apple going back t

Does anyone have any history on what motivated the industry to commoditize PC hardware? The current trend is in the opposite direction, with apple going back to proprietary components and the mobile phone/tablet market is moving further and further from user serviceable designs. Unless there's some economic driver (being able to source from more suppliers?) I don't see this happening.

JustSomeNobody 9y ago
This may help: https://en.wikipedia.org/wiki/IBM_Personal_Computer https://en.wikipedia.org/wiki/IBM_Personal_Computer
- combatentropy 9y ago
  Thanks. Shouldn't the clickbait headline to the longwinded article say "1980s"?
  - jhbadger 9y ago
    No, because it isn't really talking about the actual IBM PC (which was decidedly not a cheap commodity or modular to any real degree), but the PC clones of the 1990s which were. Although to argue with the article, while cheap and modular PCs were great for the consumer, they weren't great for the producers, ultimately leading to IBM itself exiting the PC market due to low margins. So I doubt Tesla et al, want to go down that route.
mikestew 9y ago
I don't understand your question. The original IBM PC was built using as many commercial, off-the-shelf parts as possible in order to shorten time-to-market. The hardware was commoditized from day one, by design.
- pdelbarba 9y ago
  I'm asking how commoditized hardware come to be over earlier computers with non COTS parts. If this is the first modular computer, why was all that hardware on the shelves in the first place?
  - ghaff 9y ago
    It wasn't really a new concept with the IBM PC. (Though it certainly was for IBM.) To greater or lesser degrees, a lot of the early PCs such as S-100 bus systems were largely built up using boards and peripherals from various manufacturers. For one thing, these were mostly hobbyist systems, which implied the buyers wanted to tinker (and were also relatively price-sensitive). This was in contrast to minicomputers and other similar systems (and mainframes of course), where the manufacturers pretty much designed and built everything (including the OS) and made it as difficult as possible for third-parties to play. The IBM PC was a new design so third-parties had to redesign products for it in some cases but it wasn't some sort of never-been-seen-before approach.
  - cwyers 9y ago
    It wasn't the first modular computer. The big thing with the IBM PC is that IBM didn't manage to have a single significant component that they licensed exclusively; the most obvious culprit for this was DOS, which they licensed from Microsoft. Other companies at the time had OSes they controlled the distribution of. But Microsoft was able and very willing to sell any company the rights to distribute DOS.
    - mohaine 9y ago
      Well, there was the custom BIOS that they owned outright. This was enough until it was "clean room" cloned.
  - ethbro 9y ago
    Look at the question the other way: from what at the time did not seem a COTS system from IBM, the lack of exclusive patents and sourcing rights allowed competitors to make those standards into deep facto COTS parts via market adoption of clones that also implemented them.
biofox 9y ago
I suspect it started with multiple companies creating PC clones to capture some of IBM's market, eventually leading to a defacto standard, with multiple vendors working for interoperability. I might be wrong though.
yuhong 9y ago
Even before the IBM PC there was the Altair. Ed Roberts mentioned in http://www.virtualaltair.com/virtualaltair.com/mits0011.asp http://www.virtualaltair.com/virtualaltair.com/mits0011.asp that “We should have copyrighted the name or patented the bus”.
Arizhel 9y ago
>Does anyone have any history on what motivated the industry to commoditize PC hardware? The current trend is in the opposite direction Exactly, this article is naive silliness in the extreme that fails to understand what drove standardization in the first place, and also fails to recognize the current state in the PC industry. As you said, the current trend is in the opposite direction. Open-architecture computers are mostly dead; you only still see them with hobbyist-built desktop PCs with ATX, miniITX etc. motherboards, which are a small fraction of the market now. SFF computers from places like Dell are mostly proprietary (the motherboards and chassis and power supplies are, the hard drives and memory are not). You can't swap a power supply from a SFF Dell to a SFF HP; even if the electrical connection is the same (doubtful, but it could be rigged up), the physical size just isn't compatible. Desktops are a niche industry now; most PCs sold these days are laptops, and those are highly proprietary, though they may use some standardized components like memory and WiFi cards. What drove standardization in the early 90s was the "rise of the clones": the IBM PC and its successors proved to be a good design that ran a lot of desirable software and had easily copied BIOS plus handy expansion slots, and a bunch of copycats jumped in, making compatible versions. IBM tried to kill them but failed, thankfully, and suddenly all kinds of companies were making parts for them and whole IBM-compatible computers. All the interfaces that IBM chose became standards. It didn't last though: for example, the standard floppy disks were the ones IBM chose (360k & 1.2MB 5.25", 720k & 1.44MB 3.5" later), however by the early-mid 90s when IBM tried to push to 2.88MB, no one jumped on board because they had stopped following IBM, and the industry couldn't agree on a single standard for the now-too-small 1.44MB floppies, and that's why were were stuck with those through the mid-2000s until MS finally stopped requiring them for drivers. There were a bunch of attempts at making replacements: 21MB "flopticals", 100MB Zip disks, LS-120, etc., but they were all proprietary and the industry couldn't or wouldn't choose one as an actual standard, so none of them became really ubiquitous to the point where you could expect any random recent PC to have a drive for it. Eventually the matter was settled by CD-Rs, which had their own problems (large size, and not re-writeable, until the CD-RWs). What drives standardization these days is separation of component makers from their customers, and also sheer inertia, and sometimes efforts by large companies (like Intel) to drive standardization to improve the industry which benefits themselves. Intel drove the standardization of USB and also SATA, for instance, as well as the now-decrepit ATX standard (they also tried to push the BTX standard which didn't go anywhere). But this doesn't extend to things like the form factors of power supplies, which are different once you leave ATX-land, or worse things like laptop power supplies, which are completely all over the map (different connectors, different voltages, etc.). So where you have a small number of vendors selling a component to a large number of vendors, you get standards, as we see with hard drives (both mechanical and SSD), because that makes it easier for the component makers, but for many other things (like laptop power supplies) you don't. The idea that an electric car would have interchangeable, industry-standard parts is a pipe dream. EVs aren't that different from ICE cars; it's only the powertrain that's really different. And we haven't seen much standarization in ICE cars between automakers, only within automaker lines. So you can frequently exchange brake parts between different models from a make (because it's easier for their engineering staff to reuse designs, and easier for their suppliers too), but you can't exchange them between brands that much unless maybe they're sharing suppliers. From what I've seen from several decades observing the PC industry, standardization is hard: competing companies rarely want to sit down together and choose standards, so they don't unless they have a really good reason, such as a critical component maker forces it on them because the component maker doesn't want to have dozens of different versions for all their customers, unless the customers are willing to pay extra for that (which they usually aren't).

JustSomeNobody 9y ago

This may help: https://en.wikipedia.org/wiki/IBM_Personal_Computer https://en.wikipedia.org/wiki/IBM_Personal_Computer

combatentropy 9y ago
Thanks. Shouldn't the clickbait headline to the longwinded article say "1980s"?
- jhbadger 9y ago
  No, because it isn't really talking about the actual IBM PC (which was decidedly not a cheap commodity or modular to any real degree), but the PC clones of the 1990s which were. Although to argue with the article, while cheap and modular PCs were great for the consumer, they weren't great for the producers, ultimately leading to IBM itself exiting the PC market due to low margins. So I doubt Tesla et al, want to go down that route.

mikestew 9y ago

I don't understand your question. The original IBM PC was built using as many commercial, off-the-shelf parts as possible in order to shorten time-to-market. The hardware was commoditized from day one, by design.

pdelbarba 9y ago
I'm asking how commoditized hardware come to be over earlier computers with non COTS parts. If this is the first modular computer, why was all that hardware on the shelves in the first place?
- ghaff 9y ago
  It wasn't really a new concept with the IBM PC. (Though it certainly was for IBM.) To greater or lesser degrees, a lot of the early PCs such as S-100 bus systems were largely built up using boards and peripherals from various manufacturers. For one thing, these were mostly hobbyist systems, which implied the buyers wanted to tinker (and were also relatively price-sensitive). This was in contrast to minicomputers and other similar systems (and mainframes of course), where the manufacturers pretty much designed and built everything (including the OS) and made it as difficult as possible for third-parties to play. The IBM PC was a new design so third-parties had to redesign products for it in some cases but it wasn't some sort of never-been-seen-before approach.
- cwyers 9y ago
  It wasn't the first modular computer. The big thing with the IBM PC is that IBM didn't manage to have a single significant component that they licensed exclusively; the most obvious culprit for this was DOS, which they licensed from Microsoft. Other companies at the time had OSes they controlled the distribution of. But Microsoft was able and very willing to sell any company the rights to distribute DOS.
  - mohaine 9y ago
    Well, there was the custom BIOS that they owned outright. This was enough until it was "clean room" cloned.
- ethbro 9y ago
  Look at the question the other way: from what at the time did not seem a COTS system from IBM, the lack of exclusive patents and sourcing rights allowed competitors to make those standards into deep facto COTS parts via market adoption of clones that also implemented them.

biofox 9y ago

I suspect it started with multiple companies creating PC clones to capture some of IBM's market, eventually leading to a defacto standard, with multiple vendors working for interoperability. I might be wrong though.

yuhong 9y ago

Even before the IBM PC there was the Altair. Ed Roberts mentioned in http://www.virtualaltair.com/virtualaltair.com/mits0011.asp http://www.virtualaltair.com/virtualaltair.com/mits0011.asp that “We should have copyrighted the name or patented the bus”.

Arizhel 9y ago

>Does anyone have any history on what motivated the industry to commoditize PC hardware? The current trend is in the opposite direction Exactly, this article is naive silliness in the extreme that fails to understand what drove standardization in the first place, and also fails to recognize the current state in the PC industry. As you said, the current trend is in the opposite direction. Open-architecture computers are mostly dead; you only still see them with hobbyist-built desktop PCs with ATX, miniITX etc. motherboards, which are a small fraction of the market now. SFF computers from places like Dell are mostly proprietary (the motherboards and chassis and power supplies are, the hard drives and memory are not). You can't swap a power supply from a SFF Dell to a SFF HP; even if the electrical connection is the same (doubtful, but it could be rigged up), the physical size just isn't compatible. Desktops are a niche industry now; most PCs sold these days are laptops, and those are highly proprietary, though they may use some standardized components like memory and WiFi cards. What drove standardization in the early 90s was the "rise of the clones": the IBM PC and its successors proved to be a good design that ran a lot of desirable software and had easily copied BIOS plus handy expansion slots, and a bunch of copycats jumped in, making compatible versions. IBM tried to kill them but failed, thankfully, and suddenly all kinds of companies were making parts for them and whole IBM-compatible computers. All the interfaces that IBM chose became standards. It didn't last though: for example, the standard floppy disks were the ones IBM chose (360k & 1.2MB 5.25", 720k & 1.44MB 3.5" later), however by the early-mid 90s when IBM tried to push to 2.88MB, no one jumped on board because they had stopped following IBM, and the industry couldn't agree on a single standard for the now-too-small 1.44MB floppies, and that's why were were stuck with those through the mid-2000s until MS finally stopped requiring them for drivers. There were a bunch of attempts at making replacements: 21MB "flopticals", 100MB Zip disks, LS-120, etc., but they were all proprietary and the industry couldn't or wouldn't choose one as an actual standard, so none of them became really ubiquitous to the point where you could expect any random recent PC to have a drive for it. Eventually the matter was settled by CD-Rs, which had their own problems (large size, and not re-writeable, until the CD-RWs). What drives standardization these days is separation of component makers from their customers, and also sheer inertia, and sometimes efforts by large companies (like Intel) to drive standardization to improve the industry which benefits themselves. Intel drove the standardization of USB and also SATA, for instance, as well as the now-decrepit ATX standard (they also tried to push the BTX standard which didn't go anywhere). But this doesn't extend to things like the form factors of power supplies, which are different once you leave ATX-land, or worse things like laptop power supplies, which are completely all over the map (different connectors, different voltages, etc.). So where you have a small number of vendors selling a component to a large number of vendors, you get standards, as we see with hard drives (both mechanical and SSD), because that makes it easier for the component makers, but for many other things (like laptop power supplies) you don't. The idea that an electric car would have interchangeable, industry-standard parts is a pipe dream. EVs aren't that different from ICE cars; it's only the powertrain that's really different. And we haven't seen much standarization in ICE cars between automakers, only within automaker lines. So you can frequently exchange brake parts between different models from a make (because it's easier for their engineering staff to reuse designs, and easier for their suppliers too), but you can't exchange them between brands that much unless maybe they're sharing suppliers. From what I've seen from several decades observing the PC industry, standardization is hard: competing companies rarely want to sit down together and choose standards, so they don't unless they have a really good reason, such as a critical component maker forces it on them because the component maker doesn't want to have dozens of different versions for all their customers, unless the customers are willing to pay extra for that (which they usually aren't).