In this classical view, computers are classified by size. Mainframes were large machines that filled rooms, minicomputers were smaller (the size of refrigerators), personal computers could fit on (or under) a desk, and smartphones can fit in your pocket. Servers pose a bit of a problem for this classification system, as servers are about the size of personal computers. They occupy the same "space" as personal computers. What distinguishes a server is that it can be mounted in a rack.
I have been pondering a different classification system. This new view looks beyond the hardware and the peripheral devices, and considers the major concerns of the people operating (or purchasing) computers.
Mainframes were big, but simple. The main concerns were processor speed, memory size, and storage. Storage was usually magnetic tape, but could have been magnetic disk (or drum) or possibly punch cards. Computer systems were rated based on the number of jobs they could run (either at one time or over a single night).
Minicomputers were smaller and still had processors and memory and storage, but the rating system had changed. Instead of the "number of jobs", people were now concerned with the "number of users". Minicomputer had terminals, something that mainframe computers initially lacked. (Mainframes did acquire terminals as purchasers wanted on-line systems, but the first and primary purpose of a mainframe was to run jobs efficiently.) The people who purchased minicomputers wanted to know how many users it could support.
Personal computers saw a different configuration for hardware. Computers kept processors, memory, and storage, but dropped the notion of "terminal". Personal computers were one-per-user, and instead of a terminal (a remote display and keyboard connected usually via a serial line) they used a keyboard attached to the main board and a display attached to a video board. While purchasers were interested in memory, storage, and processor speed, the main concern became video -- screen size, screen resolution, number of colors, and hardware dedicated to video.
Smartphones (and if we want, tablets) are different from personal computers physically, being smaller and portable. But it is not size that is the major issue with smartphones. Nor is the main concern the number of jobs, or the number of terminals, or the size and resolution of the screen. No, the big question in the purchase of a smartphone is battery life. How long does the battery last? How long to recharge? Screen size and weight are also factors, and some people are loyal to brands, so these are also factors. For smartphones, people think about quite different qualities than those for laptop computers.
Servers are different from all of these categories because people who buy servers want to run web sites, or (now more often) virtual machines. Purchasers want to complete requests (either on the physical server or on virtual machines on the physical server). They don't care about video -- that's for the requesting client to worry about. Thus, while servers are about the same size as personal computers, people think about them in very different ways. (In my classification system, servers are closer to mainframes than personal computers.)
What do these different concerns tell us? Well, it is interesting to see how concerns changed over time. Our interest in computers is not constant, we focus on different aspects of different classes of hardware.
Secondly, we should note that our interest in computers is not in the hardware or the software, but instead what the computer can do for us. We care more about how we use computers than how computers look or run.
Also, notice that how we store computers has changed over time, from large mainframes that are guarded in secure rooms to devices that are casually slipped into a pocket. (Much of this is a result of advances in technology.)
Now let's switch from hardware to software. Like hardware, software has changed over time, and our views of software have changed over time. Hardware definitely preceded software; indeed, the first computers were hard-wired to perform calculations and there was no software as we think of it.
Software for mainframes was designed for business purposes (accounting, inventory management, billing, etc.) and military purposes (ballistics tables were among the first applications). Software for minicomputers was designed for data analysis, which could leverage the interactivity of minicomputers with their terminals. The "killer app" for personal computers was the spreadsheet (Visicalc at first, then Lotus 1-2-3, and later Excel). For servers, databases and the web server; for smartphones, maps and GPS systems.
Our use of computers (hardware and software) has changed over time. Our expectations have changed. Initial uses of computers were "obvious": calculations better handled by machine. Later uses were not necessarily obvious but perhaps even more compelling -- including games such as "Space Invaders" and "Angry Birds".
This trend, I think, may tell us something about computing in the future. We may be more interested in computers (and applications) that serve us, that do things for us, than computers that don't. Or, we may be interested in computers that meet specific needs better than the computers we have today.
But that may not be a surprise.
Personal computers saw a different configuration for hardware. Computers kept processors, memory, and storage, but dropped the notion of "terminal". Personal computers were one-per-user, and instead of a terminal (a remote display and keyboard connected usually via a serial line) they used a keyboard attached to the main board and a display attached to a video board. While purchasers were interested in memory, storage, and processor speed, the main concern became video -- screen size, screen resolution, number of colors, and hardware dedicated to video.
Smartphones (and if we want, tablets) are different from personal computers physically, being smaller and portable. But it is not size that is the major issue with smartphones. Nor is the main concern the number of jobs, or the number of terminals, or the size and resolution of the screen. No, the big question in the purchase of a smartphone is battery life. How long does the battery last? How long to recharge? Screen size and weight are also factors, and some people are loyal to brands, so these are also factors. For smartphones, people think about quite different qualities than those for laptop computers.
Servers are different from all of these categories because people who buy servers want to run web sites, or (now more often) virtual machines. Purchasers want to complete requests (either on the physical server or on virtual machines on the physical server). They don't care about video -- that's for the requesting client to worry about. Thus, while servers are about the same size as personal computers, people think about them in very different ways. (In my classification system, servers are closer to mainframes than personal computers.)
What do these different concerns tell us? Well, it is interesting to see how concerns changed over time. Our interest in computers is not constant, we focus on different aspects of different classes of hardware.
Secondly, we should note that our interest in computers is not in the hardware or the software, but instead what the computer can do for us. We care more about how we use computers than how computers look or run.
Also, notice that how we store computers has changed over time, from large mainframes that are guarded in secure rooms to devices that are casually slipped into a pocket. (Much of this is a result of advances in technology.)
Now let's switch from hardware to software. Like hardware, software has changed over time, and our views of software have changed over time. Hardware definitely preceded software; indeed, the first computers were hard-wired to perform calculations and there was no software as we think of it.
Software for mainframes was designed for business purposes (accounting, inventory management, billing, etc.) and military purposes (ballistics tables were among the first applications). Software for minicomputers was designed for data analysis, which could leverage the interactivity of minicomputers with their terminals. The "killer app" for personal computers was the spreadsheet (Visicalc at first, then Lotus 1-2-3, and later Excel). For servers, databases and the web server; for smartphones, maps and GPS systems.
Our use of computers (hardware and software) has changed over time. Our expectations have changed. Initial uses of computers were "obvious": calculations better handled by machine. Later uses were not necessarily obvious but perhaps even more compelling -- including games such as "Space Invaders" and "Angry Birds".
This trend, I think, may tell us something about computing in the future. We may be more interested in computers (and applications) that serve us, that do things for us, than computers that don't. Or, we may be interested in computers that meet specific needs better than the computers we have today.
But that may not be a surprise.
No comments:
Post a Comment