Hardware independence has little to do with it. Before MSDOS (bought from Seattle Computer and substantially rewritten for version 2) there was CPM, which was also hardware agnostic, and S100 machines which provided a degree of bus compatibility. What vendors wanted was to be IBM compatible and MS sold them the OS that was in use on the IBM hardware.
IBM's publishing was a little more complicated than that when I was involved. Basic product manuals came with the software license and additional copies could be bought. Redbooks (experience reports written by practitioners) were distributed for a nominal cost and are massively useful (they seem to have gone up in price). There was a category of licensed manual that was hard to get (OS logic and data structure manuals and the like) that would have been available with the software license. Most of the unlicensed IBM manuals are available on the net from IBM these days.
MSDN in 1995? MS didn't have the idea, IBM did. MSDN was good if you wanted the Finnish version of NT 3.1. They didn't put tools in the early MSDN releases, just zillions of versions of MS Operating system products. The IBM Developer Connection CDs at about the same time were about the same size but consisted of CDs full of tools (compilers, databases, comms software etc) and were much more useful.
If you want to know how to use IBM software, look at the IBM redbooks. They aren't manuals, they are people from both inside and outside IBM writing about how software really works, sometimes in cookbook format, sometimes as a sort of overview (do X this way, this bit is hard to use, this bit is really good, this bit is flakey).
Actually, the power of ten thing pre-dates the retail market disks. Early IBM disks (the 3330 and the like) for example had capacity that was always expressed in powers of 1000.
Telstra Cable: AUD59.95, DOCSIS, peak bandwidth actually seen is about 5mbit (good for the odd Linux distribution), no idea what the theoretical bandwidth is, 10GB/month with bandwidth throttling after 10GB. Annoying requirement for a login client (heartbeat like RoadRunner had/has).
Optus have a similar cable offering for a similar price but with no client software requirement.
Why outsource without meaningfully tracking or managing progress? Doing this management internationally is harder than people think, even when you are part of the same company. Its VERY hard when its another company.
Why didn't you get the source? Is the IP really the property of the outsourcer? That sounds more like a software purchase than an outsource.
Acceptance test? How was that left out of the payment schedule?
It sounds like you got burned because you aren't experienced in dealing with commercial development. The outsourcer got his money and presumably met the contract conditions (delivering working software is not a condition, its a wish because its not quantifiable). Quantifiable conditions describe what bugs are, how they are tested for and how many is an acceptable number. The payment schedule must follow the quantifiable criteria.
I think its a bit more complex than that. Notes using organisations tend to fall into three categories:
those who use it for email only (these are susceptible to an Outlook migration - the UI does not inspire lots of affection)
those who use Notes for mail and simple groupware and document based applications (this lot tend to be relatively happy with it)
those whose Notes developers got a case of the usual "Notes Dev Hubris" and tried to build an ERP system (or something equally horribly complex) in Notes - this category have a management who are looking for anything that works that isn't Notes, they will buy Outlook/Exchange sometime soon.
The first group are only committed by the cost of migration and the unfortunate bits of the UI are a reason to leave.
The third group are leaving Notes.
The members of the second group are slowly converting themselves into the third group.
Although Notes is a very good product in some senses, unless the Notes developers do something about categories one and three they will become irrelevant.
Don't forget the purchase process... its not very amusing to get to the implementation of a product and have the client's admin people announce that they don't like the security of the product, the way that it installs or how it runs. You look at the admins, you look at the people who BOUGHT the thing (after due dilligence including lots of frequently off-base technical questions) and you wonder why you have to fix their internal problem. Meanwhile they are looking at you waiting for you to make the sysadmin happy.
Slashdot Mirror
Hardware independence has little to do with it. Before MSDOS (bought from Seattle Computer and substantially rewritten for version 2) there was CPM, which was also hardware agnostic, and S100 machines which provided a degree of bus compatibility. What vendors wanted was to be IBM compatible and MS sold them the OS that was in use on the IBM hardware.
I thought they used seven track drives...
IBM's publishing was a little more complicated than that when I was involved. Basic product manuals came with the software license and additional copies could be bought. Redbooks (experience reports written by practitioners) were distributed for a nominal cost and are massively useful (they seem to have gone up in price). There was a category of licensed manual that was hard to get (OS logic and data structure manuals and the like) that would have been available with the software license. Most of the unlicensed IBM manuals are available on the net from IBM these days.
MSDN in 1995? MS didn't have the idea, IBM did. MSDN was good if you wanted the Finnish version of NT 3.1. They didn't put tools in the early MSDN releases, just zillions of versions of MS Operating system products. The IBM Developer Connection CDs at about the same time were about the same size but consisted of CDs full of tools (compilers, databases, comms software etc) and were much more useful.
If you want to know how to use IBM software, look at the IBM redbooks. They aren't manuals, they are people from both inside and outside IBM writing about how software really works, sometimes in cookbook format, sometimes as a sort of overview (do X this way, this bit is hard to use, this bit is really good, this bit is flakey).
Actually, the power of ten thing pre-dates the retail market disks. Early IBM disks (the 3330 and the like) for example had capacity that was always expressed in powers of 1000.
See http://en.wikipedia.org/wiki/Kibi for the gorey details of the kilo/kibi thing and some comments about disk storage.
Disk storage has traditionally been expressed in terms of powers of ten, not powers of two. Memory is expressed in powers of two.
Telstra Cable: AUD59.95, DOCSIS, peak bandwidth actually seen is about 5mbit (good for the odd Linux distribution), no idea what the theoretical bandwidth is, 10GB/month with bandwidth throttling after 10GB. Annoying requirement for a login client (heartbeat like RoadRunner had/has).
Optus have a similar cable offering for a similar price but with no client software requirement.
Why outsource without meaningfully tracking or managing progress? Doing this management internationally is harder than people think, even when you are part of the same company. Its VERY hard when its another company.
Why didn't you get the source? Is the IP really the property of the outsourcer? That sounds more like a software purchase than an outsource.
Acceptance test? How was that left out of the payment schedule?
It sounds like you got burned because you aren't experienced in dealing with commercial development. The outsourcer got his money and presumably met the contract conditions (delivering working software is not a condition, its a wish because its not quantifiable). Quantifiable conditions describe what bugs are, how they are tested for and how many is an acceptable number. The payment schedule must follow the quantifiable criteria.
I think its a bit more complex than that. Notes using organisations tend to fall into three categories:
The first group are only committed by the cost of migration and the unfortunate bits of the UI are a reason to leave.
The third group are leaving Notes.
The members of the second group are slowly converting themselves into the third group.
Although Notes is a very good product in some senses, unless the Notes developers do something about categories one and three they will become irrelevant.
Don't forget the purchase process... its not very amusing to get to the implementation of a product and have the client's admin people announce that they don't like the security of the product, the way that it installs or how it runs. You look at the admins, you look at the people who BOUGHT the thing (after due dilligence including lots of frequently off-base technical questions) and you wonder why you have to fix their internal problem. Meanwhile they are looking at you waiting for you to make the sysadmin happy.