Article 6598 of comp.sys.ibm.pc: Path: puukko!santra!tut!enea!mcvax!uunet!husc6!rutgers!gatech!ncar!ames!pacbell!pmt1!news From: news@pmt1.UUCP (Usenet news) Newsgroups: comp.sys.ibm.pc Subject: Re: OS/2 Anyone? Summary: Its still VAPORWARE Message-ID: <427@pmt1.UUCP> Date: 4 Jun 88 18:11:26 GMT References: <1866@ncsuvx.ncsu.edu> Organization: Pacific Micro Tech, El Cerrito, CA Lines: 131 In article <1866@ncsuvx.ncsu.edu>, sam@ncsuvx.ncsu.edu (Whad Upp) writes: > > > I just got back from the announcement of the Model 70 and Model 50Z PS/2s > given by IBM. They told us that OS/2 was quickly becoming THE OS standard, etc. > > I have OS/2 running, but the only application I have is a port of MicroEmacs > that I did myself. I have talked with a handfull of other people that are > in about the same position as I am. > > Is anyone doing any OS/2 development? Are people waiting for the PM to be > released? How about public domain software for OS/2? Is it out there in one > place? > > I am willing to setup an FTPable library of public OS/2 software if there > is a need. But is there? > The problem is much deeper that just waiting for presentations manager. OS/2 is to support large corporate networks using a new communications manager. The OS/2 applications on individual workstations are to be connected by a lan to communications manager. So many of OS/2s first users will be IBMs largest customers. They buy PCs by the thousands or tens of thousands. Consider the plight of a major corporation trying to implement new applications and networks under OS/2 and it is no wonder that things are going so slowly. I have been doing consulting for a large fortune 500 company which is known to be a large IBM account. They want to deploy a huge network and a large number of workstations using PS/2s and OS/2. And of course they want to buy everything from Big Blue. The project has been slipping from day one, and the major reason waiting for the new hardware and OS/2 EE. For instance: We've been waiting for the new communications manager, the one that is to allow us to consolidate our existing networks. OS/2 at this stage is not very useful, and we have to convert existing applications to use it, or wait for other vendors to do conversions. Many third party vendors are taking a wait and see approach. We can't run existing applications in the compatiblity box. We require host communications and the compatibility box switches from real to protected mode which disrupts communications. We're waiting for the lan manager so we can hook application workstations to the network when it becomes available. We're waiting for presentation manager. But since all our existing applications use other windowing software we don't know exactly where windows (sorry, Presentation Manager) fits. It is unclear that PM really has all of the tools to replace the application specific window managers, such as adequate editing and forms handling abilities. The IBM PS/2 model 50 and newly announced variants all seem too expensive for many many thousands of application workstations. Certainly far more expensive than clustering users on Unix systems (say 3-4 user per machine) and hooking the multi-user systems into the lan and network. If using OS/2 we would prefer diskless workstations which can boot from a common file server. While the model 25 is not right for us (does not run OS/2), the model 50 is not diskless. True we read about experimental diskless workstations but, IBM seems to be quiet about them. When (when ?) we get diskless PCs and are able to boot from a file server, the token ring lan appears to be too slow for the load at the current bandwidth. We hear about 16MB lan cards of the future...but where are they? So we had another thought ... lets put our applications under Dos and move them to OS/2 later. This is an attempt to end the waiting game, so we can work on our application backlog and keep our users happy. One problem: the workstations we choose for DOS may not be the correct configuration for OS/2 later on. It appeas that as OS/2 comes nearer, the workstations are being revised so they will be appropriate for the new operating system. That is just what happened with UNIX in 1982, when the first vendors told us it worked just fine without a hard disk, and that it only needed 340K to work properly. New versions of the existing Unix computers appeared within a year. The other problem with DOS is the 640K memory barrier. When you load a PC with DOS and lan manager, you lose lots of memory. Especially if you don't have a network gateway and you also have to add 3270 software in the PC. Ram disk is the obvious choice to ease our memory problems. But we're told by IBM that ram disk is incompatable with newtwork operation. You can lose interrupts when using ram-disk and this is incompatible with doing lan transmissions and handling unsolicited inbound messages. So our most functional applications are squeezed for memory under DOS. We keep reading in the trade press about DOS 3.4, but it was not announced June 2nd as rumored. IBM seems to be justifying OS/2 for applications which need to break the 640K barrier. So is may be that even if DOS 3.4 can break the 640K barrier, it may never be released, because it could remove the obvious reason for using OS/2. We can't use OS/2 because the pieces are not in place and it needs time to mature. We can't use DOS without concern that our existing applications can't grow, and with inadequate memory, performance suffers. Migrating from DOS now to OS/2 later means we risk having to retrofit or replace thousands of workstations over a large geographical area. An with DOS we still need to look for a third party gateway vendor. LU 6.2 has been pretty slow to reach critical mass too. And we wanted to buy from Big Blue. So I think a good part of the reason OS/2 is moving slowly is that there are MANY new pieces and they don't fit together right now. Most standalone workstation users will prefer to stay with DOS due to low price and availavbility of applications. OS/2 workstations are going to take 3-5MB of memory. So just having OS/2 or getting Presentation Manager may not make the market really take off. OS/2 today is likely to appeal for large scale distributed processing and big corporate networks. At least to start out. Many of the candidates for OS/2 are likely to be IBM's best customers - and they are bogged down with problems such as those discussed above. Its going to be slow going no matter what IBM or the trade journals say.