The real cost of upgrading is not the replacing of the software and hardware, but of the data.
If ye have stable systems running DOS or OS/2 or Windows 3.1 or Linux, the replacing of the software is a minimal expense to data conversion. This is what really limits Microsoft's price expansion.
It takes me about two to three hours to bring a new or reformatted box up to some sort of usable form. On the other hand, to get my home-grown data to work might take several hours. Why?
Because the batch files &c have expectations of being run from certian locations. Because files that import graphics have to be pointed at the new location, and because the clean sweep gives some room to redesign the computer.
If ye have a computer that is chugging away, doing the same mindless task without error, there is little incentive to replace. This is why databases support ancient formats: most of the money is tied up just there.
The reason i keep installing my ancient Win16 word processor Amipro 3.1 is because all my files are in that format. I have more modern word processors, but Amipro 3.1 is more "open" to text scripts that i have built for it, and is at least adequate for its function that i have not decided to replace.
Microsoft have of course, the "legacy trap". By tying your precious data to their priority formats, you are tying yourself to their software. In the case of OS/2 or Linux or whatever, you are still linked to some other thing, the good thing is that IBM is still supporting it (along with a lot of OS folk: Daniela Engbert's drivers are widely regarded).
Would not relevance of neighbouring objects come into effect. I mean, there's nothing out there with pluto, but there are lots of stuff hanging around the asteroid belt, of which ceres happens to be the largest...
It's not absolute size but importance in the neighbourhood, I would have thought....
Once the protocols stabalise, at least into a robust commercial form, I can't see why it can't be any less a long term career than, say banking.
I mean, banks bring out new products all the time, and even the core market can be somewhat changing, eg by deregulation of the market, etc. But it still does not prevent one forging a career in banking.
Likewise, once the protocols stabalise, it does not prevent people persuing a career in IT. Even major paradigm shifts like windows/unix is seen elsewhere in industry (eg metric conversion), and such do not drive massive numbers of people out of the affected industry.
The dot-com bust is probably a grander scale event similiar to the typical building booms and busts, and the massive railway boom/bust that left a large amount of infrastructure to recycle. Some UK cities have abandoned railway terminii, from defunct companies whose assets were assimilated and rationalised into more favourable terminals.
So what we have gone through in the last 20 years, and maybe the next 10, are just the sorting out of the technological structure, and we should start seeing computers stable enough for long-term careers to be forged.
the number has too many digits for even an international telephone number. So i suspect some humour.
The number is given as being a famous number, being 2^64.
On the other hand, if you're from the USA, then dialing short+short+short+short and ask the operators for connections is the way you call people (see "operator" song by Elvis).
The author ofthe number is not using the british system, but the European system. This is evident from the use of -illards, the absence of 'and', and a few other things.
The selected number here is 2^64, the decimal expansion of which is 18 446 744 073 709 551 616.
USA = eighteen pentillion, four hundred forty six quadrillion, seven hundred forty four trillion, seventy three billion, seven hundred nine million, five hundred fifty one thousand six hundred sixteen.
EUR = eighteen trillion, four hundred forty six billiard, seven hundred forty four billion, seventy three milliard, seven hundred nine million, five hundred fifty one thousand six hundred sixteen.
UK = eighteen trillion, four hundred and forty-six thousand, seven hundred and forty-four billion, seventy-three thousand, seven hundred and nine million, five hundred and fifty-one thousand six hundred and sixteen.
NEW = three hundred and sixtynine billion, one cention, fiftyseven thousand, thirtyone hundred and eleftynine million, fifty centions, seven thosand, ninety hundred and sixteen.
Can you imagine this in the hands of the corporate people who want to "cover" their paper tracks - I shudder to think this is what is becomming of the world
Warp 4 did appear on the cover of APC Magazine about a year ago. Full thing, no time lock or anything. Netscape 4.04 was there, but the service pack it needed wasn't.
It's interesting to note the comment about "by 1994 or 1995, nearly all cdroms were IDE. This is not true. A lot of cdroms at this time worked through sound cards (eg SB16). IDE drives may have *started* to appear at that time, but they were by no means common in the market place.
I too had to hunt down drivers for my cdrom, but I never had any "reinstall without concent".
The other thing is that the poster may have done an "express install", which does reformat c:. This is stated in the installation guide.
Usability has much to do with the interface as fitness for purpose. That is, the user should be able to understand what the tool does, and have some measure of control over its use.
Windows has a nice point+grunt interface, and a few tasks are relatively simple and obvious. On the other hand, lots of things are hidden in unusual places, which makes using it similar to playing KingsQuest.
The biggest fault is that there is some effort to "dumb-down" the interface, which means that guessing what to search for in the help files is quite amusing.
Windows, including gui programs, are scriptable from the command line, and so forth. But you would be lucky to find the command line options for your self unless you want to "read the binaries". For example, when starting the internet connection, I start the fire wall, the pop-up killer and then the logon. The firewall has some useful options (for closing the gui), which are not documented.
The best prize goes to Clippy. Even if you manage to find the right terms, clippy will direct you to a variety of different unrelated screens. It's not clear how certian features are implemented in documents, and Word is so bletcherous at formatting anyway, so you are reduced into "wordpad" style formatting in your 1000$-plus word-processor.
User friendliness is more to do with keeping with conventions, making sure that your tool will do what it says it will do, and drawing on past experiences. Controls that go together should be placed together. But even here, one needs to place some controls so that they are not accidently purged (eg ctrl+A = select all, ctrl+Z zapp all would be bad.)
The OS/2 user interface is less user-friendly to point-and-grunt but is more scriptable. Unfortunately, there is no obvious way to connect point+grunt actions to the rexx interface: specifically, one can not give a rexx name to a hand-made icon.
The whole point of the UNIX-style pipe is that there was an obvious user model of what the tools did, and tools just operated in an obvious way. Even here, there are problems when the -? option is put to STDERR (so "x -? > x.help" does not work).
I think part of the problem is that there is no standard GUI patterns in the OSS world, so everyone's doing their own thing. Maybe we need to look at some sort of "how your program should interact with the gui" thing for OSS.
In Windows and OS/2, this is relatively clear. Not perfect, but passable enough to make their interfaces something people want to use.
While Larrikin is usually glossed as an unruly person, over here it carries aslo the sort of connetations of the innocent mischief young adults get up to. Something like a "hood", I suppose. A lot of our words do not translate to american, since ours carry shades of meaning not present in the American.
Moon (v intrans) means to show the exposed backside, often in a reverse bow. It sort of looks like the moon, if the pants are dropped to mid thigh. Don't forget, we see the moon right-way up, not upside down as you do in the Northern hemisphere.
"larrikin moon" then translates to a moon as a harmless mischief. Something about three orders up to giving what the americans use the "finger".
Until recently, my main machine was a 486/66 with 20 meg of ram. On this, I ran some 20 operating systems.
One of these was a version of OS/2, complete with gui, 4os2 and cdburner, that lived in 10 megabytes.
The installation was not hard - sysintx and rar did it.
So I could use the main version of OS/2 without having to worry about chewing up memory for the cd-burner. OS/2 breaks the 504 barrier for HPFS partitions. So does Linux, and Windows NT, the former installed but never booted.
OS/2 has a considerably smaller footprint, given that a lot of it can be installed on another drive.
The idea of having a small footprint is not bad at all. You can make a boot cd that runs the desired OS off the cdrom and ramdisk. This is how eComStation installs.
In fact, the notion of one OS for all tasks is quite unecconomical, especially if the machine is to run unattended, or in a specific activity.
You can burn cdroms off in 20 mb of ram. There are utilities that unload dlls to expediate the process.. (eg allocmem)
Whether the use of RPN over algebraic calculators will make your life easier is a moot point. It's the same issue as Qwerty vs Dorvak keyboards.
I am not sure about your remark about RPN users being more arithmetically challanged.
RPN is an easy process to implement in code. It is also very useful when you can't pass parameters in a call. This is because an operation like + or * finds the inputs in the stack, and leaves the stack in a known condition. As a result, *only one operator at a time* is in use.
If you are using something that has an open arena, or variable pool [such as BASIC or REXX], then because only one operator is active, *all operators* can use the same internal variables.
If the stack is implemented as an array [A, L, X, Y, Z, T, pi, memory], one can implement storage and recall operators by pointing the pointer at memory.
If you think that RPN is something to do with post operators, eg 5,3+, there are differences between the commercial 4-deep implementations of RPN and the infinite stack.
But basically, RPN is just a different way of doing things.
The issue is not so much that that extra features have been added, but that the intent is not correctly communicated, or is inappropriate
For example, the WPS applied to the OS/2 desktop is a wonderous thing, one that people desire in other systems. When this file-viewing device is applied to files in general [eg DRIVES], the result is a nightmare. Drives is *not* one of os/2's better features.
Windows copied this feature into their shell, along with a network browser. Unlike the OS/2 one, these ones *can not be hidden*, especially without corrupting the operating system. [deleting Network Neighbourhood removes UNC support].
It's not that the "start menu" is totally bad either. It relies on an established practice of menus. So does the send-to [as a configurable context menu that allows drag-and-drop to otherwise hidden targets]. Folders = submenus. So you can have submenus in the send-to as well.
It's not that one can't make the windows shell liveable. Create a directory called grotto, and move these folders from Windows: sendto, start menu, desktop, shellnew, recent. You can create other folders there as well.
Create an icon with the command explorer.exe/e,/root=c:\grotto,start menu
This gives you a super-program manager that you can fix your start menu, send-to, etc, as well as drag out recently edited docs for shortcuts to the desk.
The other issue of what happens whens when one closes dialogs (as to whether it's an OK or Cancel), frustrates users to no end.
The issue is not so much as Cruft, but the lack of consistancy. Were cars like this, they would be hazardous.
RPN calculators have a lot less "pending" operations than algebraic calculators. This is because all of the argurments of any function are present before the operation is done.
The actual control logic takes about 10 lines of code, and has very few pending values [just the stack]. Further more, the RPN logic needs to know nothing about what's in the stack (eg matricies, complex numbers).
One is not dependent on the manufactures implementation of pending operations. In an algebraic calculator 3*4+2*5 can give all sorts of different values, eg 22, 70. The same command in RPN is 3~4*2~5*+ (~ is enter) alwaus gives 22.
This lopks strange, but is the exact way you would do the calculation yourself: get 3, get 4, multiply. Get 2, get 5, multipy. Add the two together.
Of course, algebraic calculators are not strictly algebraic, eg cos 60 is entered as 60 cos in both systems.
In practice, the last time I looked, RPN was doing quite well with the financial crowd, since both it and tape-calculators (ie += -= logic) take the operator after the number. That is, to add 5, one goes 5 + or 5 +=, rather than + 5.
If one is used to using prefix-operators, you will find the algebraic form easier and faster. If you find the postfix-operators, you will find RPN and Strip-adders easier to use. If you normally expect people to be able to use your calcualtor, you should have both kinds at your desk.
The thing that tends to cost money is not "open source" vs "close source", but "integrated" vs "cobbled together".
It is certianly cheaper and easier to buy integrated as opposed to cobble a solution together yourself. What happens is when the integration becomes unstuck.
When I bought XTREE 2.5, it came with integrated zip integration. This allowed one to look inside and work with zip files as if they were directories. Of course PKZIP 2 then came out with a new format, which xtree could not deal with.
When I bought ZTree, it used external versions of these programs. So the unpacker for zip files in ztree is pkzip2 itself, not some some internal routine. So while ztree knows about zip and rar and cab files, actually opening these means that unzip, unrar, uncab must be available.
The are advantages and disadvantages to both approaches. If I go for Norton Commander or FC/2, then Norton Commander has *its* set of inbuilt conversions, while FC/2 uses the same set of utilities that Ztree uses.
Of course, it goes on and on. Word + Windows is a more expensive deal than WP5.1, but the printer drivers live in Windows, not WP5.1, and so Word will have printer drivers long after WP 5.1. Also, Excel can share the same printer drivers and fonts, whereas Lotus has a different printer-driver. Up goes bloat, up goes cost, down goes upgrade-costs.
The net effect is that it initially costs more to buy and install a component system, against an integrated package, but the long term maintenance is less, if it is done properly.
While it is more expensive, both in time and money, to do things in peices, the results are infinitely more flexiable. This may suit the home hacker, but for most office workers, this flexiability adds confusion, rather than confort.
Each process costs more, depending on what you choose to value.
Plain text is the simplest format - but it can not handle bolding &c. On the other hand, html handles bolding and cross-reference, but can come with all sorts of nasties.
Even delicate formatting does not chew up a lot of space: TeX is a classic example of delicate formatting + small size.
The problem is that we have bloated formats, that not only preserve the document, but the user's printer and last view of it.
We have all sorts of fun trying to print documents sent to us by users who printed to a bypass tray last! [word preserves that information].
A limit gets people into thinking about small file formats, prehaps. At least not using excessive ransom fonts &c.
But if a megabyte stores only four documents where it used to store 40, then this is going to increase storage. Some storage and pipe issues have been addressed: faster, bigger drives, faster networking. But some remain stuck where it was 10 years ago: dial-up modems and floppy disks.
I suppose that it is as much the fault of the software manufacturers problem, since their bloated formats (with no option to make it smaller), is crushing the infrastructure.
Hon. The OS/2 community is quite alive, even if IBM is not pushing OS/2 at them.
You forget, the browser bundled with OS/2 4.x (the on-ramp to the internet) is webexplorer. It does alright as a front-end to pages on a cdrom.
It was not the first with USB, but IBM added support for USB in fix-pack 40 or something.
Ok, OS/2 may be jerry-rigged. So is Windows, Linux, BeOS. Why. So that people can run jerry-rigged apps. I mean, the investment in computers is not the OS but the apps. OS's run them, and they acquire a fair bit of crud over time. Just the way the world works, hon
On the other hand, OS/2 was the first really big prime-time OS to hit the market. It did so in 1992. At that stage, internet was not a big thing, and mosaic (ie webexplorer) was bleeding edge browser. Ditto tcp/ip.
Hon. Please learn the truth before wrangling with me.
IBM were a convicted monopolist long before OS/2 appeared on the scene. IBM is a hardware company who make software. Their software is none the less good.
For being such a loser product, it does a pretty good attempt of staying alive. Its community is every bit as vibrant as that of Linux, despite being a closed source operating system. Why. It was the firstest with the mostest. OS/2 v 3 now has USB support. Hmm. Not Windows NT 4.
IBM also complied with the antitrust people.
It must never be forgotten that the roots of Linux lies in yet another anti-trust case: that of AT&T. They developed unix, and the source code for unix was released under assorted agreements and restrictions of antitrust. This became the foundation of the open-source stuff.
On the other hand, MS tightened the screws in every way they are legally allowed to, such as using massive reserves to literally capture markets with loss-leaders. I do not trust MSFT.
Slashdot Mirror
If ye have stable systems running DOS or OS/2 or Windows 3.1 or Linux, the replacing of the software is a minimal expense to data conversion. This is what really limits Microsoft's price expansion.
It takes me about two to three hours to bring a new or reformatted box up to some sort of usable form. On the other hand, to get my home-grown data to work might take several hours. Why?
Because the batch files &c have expectations of being run from certian locations. Because files that import graphics have to be pointed at the new location, and because the clean sweep gives some room to redesign the computer.
If ye have a computer that is chugging away, doing the same mindless task without error, there is little incentive to replace. This is why databases support ancient formats: most of the money is tied up just there.
The reason i keep installing my ancient Win16 word processor Amipro 3.1 is because all my files are in that format. I have more modern word processors, but Amipro 3.1 is more "open" to text scripts that i have built for it, and is at least adequate for its function that i have not decided to replace.
Microsoft have of course, the "legacy trap". By tying your precious data to their priority formats, you are tying yourself to their software. In the case of OS/2 or Linux or whatever, you are still linked to some other thing, the good thing is that IBM is still supporting it (along with a lot of OS folk: Daniela Engbert's drivers are widely regarded).
1.. Date of article = 1 April 2.. 141569 is pretty close to the digits of pi, pi = 3.14159265359 Might be coincidene...... :)
On the other hand, if Jupiter is reckoned as a very dull star, then is Europa one of its planets...
Would not relevance of neighbouring objects come into effect. I mean, there's nothing out there with pluto, but there are lots of stuff hanging around the asteroid belt, of which ceres happens to be the largest...
It's not absolute size but importance in the neighbourhood, I would have thought....
I mean, banks bring out new products all the time, and even the core market can be somewhat changing, eg by deregulation of the market, etc. But it still does not prevent one forging a career in banking.
Likewise, once the protocols stabalise, it does not prevent people persuing a career in IT. Even major paradigm shifts like windows/unix is seen elsewhere in industry (eg metric conversion), and such do not drive massive numbers of people out of the affected industry.
The dot-com bust is probably a grander scale event similiar to the typical building booms and busts, and the massive railway boom/bust that left a large amount of infrastructure to recycle. Some UK cities have abandoned railway terminii, from defunct companies whose assets were assimilated and rationalised into more favourable terminals.
So what we have gone through in the last 20 years, and maybe the next 10, are just the sorting out of the technological structure, and we should start seeing computers stable enough for long-term careers to be forged.
is eParliment eHouse, eCanberra, e2600 :)
the number has too many digits for even an international telephone number. So i suspect some humour.
The number is given as being a famous number, being 2^64.
On the other hand, if you're from the USA, then dialing short+short+short+short and ask the operators for connections is the way you call people (see "operator" song by Elvis).
The selected number here is 2^64, the decimal expansion of which is 18 446 744 073 709 551 616.
USA = eighteen pentillion, four hundred forty six quadrillion, seven hundred forty four trillion, seventy three billion, seven hundred nine million, five hundred fifty one thousand six hundred sixteen.
EUR = eighteen trillion, four hundred forty six billiard, seven hundred forty four billion, seventy three milliard, seven hundred nine million, five hundred fifty one thousand six hundred sixteen.
UK = eighteen trillion, four hundred and forty-six thousand, seven hundred and forty-four billion, seventy-three thousand, seven hundred and nine million, five hundred and fifty-one thousand six hundred and sixteen.
NEW = three hundred and sixtynine billion, one cention, fiftyseven thousand, thirtyone hundred and eleftynine million, fifty centions, seven thosand, ninety hundred and sixteen.
Not only that, but they worked in a mine, so this theme could be replicated.
Now, if there is a Snow White, or a wicked witch of the West.....
Can you imagine this in the hands of the corporate people who want to "cover" their paper tracks - I shudder to think this is what is becomming of the world
Suppose that some "public interest" suggestion could be put to bear on MS acquiring companies in related fields....
Suppose prowling around ships that lie in traffic lanes is more dangerous and hazardous than prowling around abandoned railway stations.
In any case, industrial archeology is an interesting study. Look, don't take.
Warp 4 did appear on the cover of APC Magazine about a year ago. Full thing, no time lock or anything. Netscape 4.04 was there, but the service pack it needed wasn't.
I too had to hunt down drivers for my cdrom, but I never had any "reinstall without concent".
The other thing is that the poster may have done an "express install", which does reformat c:. This is stated in the installation guide.
Usability has much to do with the interface as fitness for purpose. That is, the user should be able to understand what the tool does, and have some measure of control over its use.
Windows has a nice point+grunt interface, and a few tasks are relatively simple and obvious. On the other hand, lots of things are hidden in unusual places, which makes using it similar to playing KingsQuest.
The biggest fault is that there is some effort to "dumb-down" the interface, which means that guessing what to search for in the help files is quite amusing.
Windows, including gui programs, are scriptable from the command line, and so forth. But you would be lucky to find the command line options for your self unless you want to "read the binaries". For example, when starting the internet connection, I start the fire wall, the pop-up killer and then the logon. The firewall has some useful options (for closing the gui), which are not documented.
The best prize goes to Clippy. Even if you manage to find the right terms, clippy will direct you to a variety of different unrelated screens. It's not clear how certian features are implemented in documents, and Word is so bletcherous at formatting anyway, so you are reduced into "wordpad" style formatting in your 1000$-plus word-processor.
User friendliness is more to do with keeping with conventions, making sure that your tool will do what it says it will do, and drawing on past experiences. Controls that go together should be placed together. But even here, one needs to place some controls so that they are not accidently purged (eg ctrl+A = select all, ctrl+Z zapp all would be bad.)
The OS/2 user interface is less user-friendly to point-and-grunt but is more scriptable. Unfortunately, there is no obvious way to connect point+grunt actions to the rexx interface: specifically, one can not give a rexx name to a hand-made icon.
The whole point of the UNIX-style pipe is that there was an obvious user model of what the tools did, and tools just operated in an obvious way. Even here, there are problems when the -? option is put to STDERR (so "x -? > x.help" does not work).
I think part of the problem is that there is no standard GUI patterns in the OSS world, so everyone's doing their own thing. Maybe we need to look at some sort of "how your program should interact with the gui" thing for OSS.
In Windows and OS/2, this is relatively clear. Not perfect, but passable enough to make their interfaces something people want to use.
I never hung around with the crowd that did this sort of thing :)
Maybe they don't see the moon that far north, so I would not be supprised.
While Larrikin is usually glossed as an unruly person, over here it carries aslo the sort of connetations of the innocent mischief young adults get up to. Something like a "hood", I suppose. A lot of our words do not translate to american, since ours carry shades of meaning not present in the American.
Moon (v intrans) means to show the exposed backside, often in a reverse bow. It sort of looks like the moon, if the pants are dropped to mid thigh. Don't forget, we see the moon right-way up, not upside down as you do in the Northern hemisphere.
"larrikin moon" then translates to a moon as a harmless mischief. Something about three orders up to giving what the americans use the "finger".
One of these was a version of OS/2, complete with gui, 4os2 and cdburner, that lived in 10 megabytes.
The installation was not hard - sysintx and rar did it.
So I could use the main version of OS/2 without having to worry about chewing up memory for the cd-burner. OS/2 breaks the 504 barrier for HPFS partitions. So does Linux, and Windows NT, the former installed but never booted.
OS/2 has a considerably smaller footprint, given that a lot of it can be installed on another drive.
The idea of having a small footprint is not bad at all. You can make a boot cd that runs the desired OS off the cdrom and ramdisk. This is how eComStation installs.
In fact, the notion of one OS for all tasks is quite unecconomical, especially if the machine is to run unattended, or in a specific activity.
You can burn cdroms off in 20 mb of ram. There are utilities that unload dlls to expediate the process.. (eg allocmem)
.
RPN is an easy process to implement in code. It is also very useful when you can't pass parameters in a call. This is because an operation like + or * finds the inputs in the stack, and leaves the stack in a known condition. As a result, *only one operator at a time* is in use.
If you are using something that has an open arena, or variable pool [such as BASIC or REXX], then because only one operator is active, *all operators* can use the same internal variables.
If the stack is implemented as an array [A, L, X, Y, Z, T, pi, memory], one can implement storage and recall operators by pointing the pointer at memory.
If you think that RPN is something to do with post operators, eg 5,3+, there are differences between the commercial 4-deep implementations of RPN and the infinite stack.
But basically, RPN is just a different way of doing things.
The issue is not so much that that extra features have been added, but that the intent is not correctly communicated, or is inappropriate
For example, the WPS applied to the OS/2 desktop is a wonderous thing, one that people desire in other systems. When this file-viewing device is applied to files in general [eg DRIVES], the result is a nightmare. Drives is *not* one of os/2's better features.
Windows copied this feature into their shell, along with a network browser. Unlike the OS/2 one, these ones *can not be hidden*, especially without corrupting the operating system. [deleting Network Neighbourhood removes UNC support].
It's not that the "start menu" is totally bad either. It relies on an established practice of menus. So does the send-to [as a configurable context menu that allows drag-and-drop to otherwise hidden targets]. Folders = submenus. So you can have submenus in the send-to as well.
It's not that one can't make the windows shell liveable. Create a directory called grotto, and move these folders from Windows: sendto, start menu, desktop, shellnew, recent. You can create other folders there as well.
Create an icon with the command /e,/root=c:\grotto,start menu
explorer.exe
This gives you a super-program manager that you can fix your start menu, send-to, etc, as well as drag out recently edited docs for shortcuts to the desk.
The other issue of what happens whens when one closes dialogs (as to whether it's an OK or Cancel), frustrates users to no end.
The issue is not so much as Cruft, but the lack of consistancy. Were cars like this, they would be hazardous.
The actual control logic takes about 10 lines of code, and has very few pending values [just the stack]. Further more, the RPN logic needs to know nothing about what's in the stack (eg matricies, complex numbers). One is not dependent on the manufactures implementation of pending operations. In an algebraic calculator 3*4+2*5 can give all sorts of different values, eg 22, 70. The same command in RPN is 3~4*2~5*+ (~ is enter) alwaus gives 22. This lopks strange, but is the exact way you would do the calculation yourself: get 3, get 4, multiply. Get 2, get 5, multipy. Add the two together.
Of course, algebraic calculators are not strictly algebraic, eg cos 60 is entered as 60 cos in both systems.
In practice, the last time I looked, RPN was doing quite well with the financial crowd, since both it and tape-calculators (ie += -= logic) take the operator after the number. That is, to add 5, one goes 5 + or 5 +=, rather than + 5.
If one is used to using prefix-operators, you will find the algebraic form easier and faster. If you find the postfix-operators, you will find RPN and Strip-adders easier to use. If you normally expect people to be able to use your calcualtor, you should have both kinds at your desk.
It is certianly cheaper and easier to buy integrated as opposed to cobble a solution together yourself. What happens is when the integration becomes unstuck.
When I bought XTREE 2.5, it came with integrated zip integration. This allowed one to look inside and work with zip files as if they were directories. Of course PKZIP 2 then came out with a new format, which xtree could not deal with.
When I bought ZTree, it used external versions of these programs. So the unpacker for zip files in ztree is pkzip2 itself, not some some internal routine. So while ztree knows about zip and rar and cab files, actually opening these means that unzip, unrar, uncab must be available.
The are advantages and disadvantages to both approaches. If I go for Norton Commander or FC/2, then Norton Commander has *its* set of inbuilt conversions, while FC/2 uses the same set of utilities that Ztree uses.
Of course, it goes on and on. Word + Windows is a more expensive deal than WP5.1, but the printer drivers live in Windows, not WP5.1, and so Word will have printer drivers long after WP 5.1. Also, Excel can share the same printer drivers and fonts, whereas Lotus has a different printer-driver. Up goes bloat, up goes cost, down goes upgrade-costs.
The net effect is that it initially costs more to buy and install a component system, against an integrated package, but the long term maintenance is less, if it is done properly.
While it is more expensive, both in time and money, to do things in peices, the results are infinitely more flexiable. This may suit the home hacker, but for most office workers, this flexiability adds confusion, rather than confort.
Each process costs more, depending on what you choose to value.
Plain text is the simplest format - but it can not handle bolding &c. On the other hand, html handles bolding and cross-reference, but can come with all sorts of nasties.
Even delicate formatting does not chew up a lot of space: TeX is a classic example of delicate formatting + small size.
The problem is that we have bloated formats, that not only preserve the document, but the user's printer and last view of it.
We have all sorts of fun trying to print documents sent to us by users who printed to a bypass tray last! [word preserves that information].
A limit gets people into thinking about small file formats, prehaps. At least not using excessive ransom fonts &c.
But if a megabyte stores only four documents where it used to store 40, then this is going to increase storage. Some storage and pipe issues have been addressed: faster, bigger drives, faster networking. But some remain stuck where it was 10 years ago: dial-up modems and floppy disks.
I suppose that it is as much the fault of the software manufacturers problem, since their bloated formats (with no option to make it smaller), is crushing the infrastructure.
You forget, the browser bundled with OS/2 4.x (the on-ramp to the internet) is webexplorer. It does alright as a front-end to pages on a cdrom.
It was not the first with USB, but IBM added support for USB in fix-pack 40 or something.
Ok, OS/2 may be jerry-rigged. So is Windows, Linux, BeOS. Why. So that people can run jerry-rigged apps. I mean, the investment in computers is not the OS but the apps. OS's run them, and they acquire a fair bit of crud over time. Just the way the world works, hon
On the other hand, OS/2 was the first really big prime-time OS to hit the market. It did so in 1992. At that stage, internet was not a big thing, and mosaic (ie webexplorer) was bleeding edge browser. Ditto tcp/ip.
IBM were a convicted monopolist long before OS/2 appeared on the scene. IBM is a hardware company who make software. Their software is none the less good.
For being such a loser product, it does a pretty good attempt of staying alive. Its community is every bit as vibrant as that of Linux, despite being a closed source operating system. Why. It was the firstest with the mostest. OS/2 v 3 now has USB support. Hmm. Not Windows NT 4.
IBM also complied with the antitrust people.
It must never be forgotten that the roots of Linux lies in yet another anti-trust case: that of AT&T. They developed unix, and the source code for unix was released under assorted agreements and restrictions of antitrust. This became the foundation of the open-source stuff.
On the other hand, MS tightened the screws in every way they are legally allowed to, such as using massive reserves to literally capture markets with loss-leaders. I do not trust MSFT.