Windows 3.1 @ 20

[CoffeeFiend]

... in the 1st world. Here in Brazil, as a rule-of-thumb, treble all the prices you see on e-Bay or amazon. And the cost-of-living is much higher than in the US/Canada, at least in the big cities, with lower wages generally. So, probably further 3-5 years will be required before it's generally affordable. And AFAIK it's even worse elsewhere.

Ouch. Here for $500 you could get a i5 2500K (can be overclocked), a decent Z68 motherboard (e.g. Gigabyte GA-Z68AP-D3), 16GB of DDR3 1600 (4x4GB) and a Radeon HD 6750 1GB... That would be equal to 50h of work at minimum wage (gross not net) which is really not that bad if you're working for more than minimum wage. I can't think of an easy way to easily compare cost of living though. I wonder how many hours it would take at minimum wage for the same upgrade kit (a whole year?)

Edit: I guess I could use GDP per capita for a simple comparison. Roughly $50000 for Canada vs $10000 for Brazil. And you say the same parts cost 3x the price there too?

[dencorso]

Well, what I'll say next is far from perfect, but serves to give you a fair enough picture of it:

Exchange Rate (Apr 5, 2012) US$1.00 = R$ 1.89

Minimum Wage (net) for Brazil: US$329.00 (per month);

Then, there's the Big Mac Index: July 2011 and update (Jan 2012). And there is the Alternative Big Mac Purchasing Power Index (which I think wasn't updated ever since). Then, there's the Gini Coefficient: some newspapers were commemorating Brazil's (0,5304 !!!) had reached in 2010 its lowest value in 50 years (portuguese)...

Some states do have higher Minimum Wages defined by State Laws, so let's take Sao Paulo State as the reference:

Minimum Wage (average, net) for Sao Paulo State: US$370.00 (per month; it's defined by type of work, but varies just ±US$5,00 among them);

Furthermore Sao Paulo City ranked 10th place (2011, it was 21st in 2010) in Mercer's Most Expensive Cities index.

[bphlpt]

In the US, when you talk about minimum wage, it's usually expressed in US$/hour gross, ie before withholding taxes, Social Security, insurance, etc. According to http://www.dol.gov/dol/topic/wages/minimumwage.htm the current US federal minimum wage is "$7.25 per hour effective July 24, 2009" gross, though it can vary from state to state ($5.15 - $9.04) So when you say "Minimum Wage (net) for Brazil: US$329.00", that's not US $ per hour obviously. But if that is net US$/week, at 40 hours per week, then net US$329.00 is equivalent to net US $8.23/hour, which looks pretty good if I understood you correctly. But of course if the cost of living is three times what it is in the US, food, housing, gas, PC parts, etc, then that's not so good. But at any rate, it seems wee need some clarification of net vs gross and time frame and verification that we're talking US $ to be sure were comparing apple to apples as best we can. The Big Mac Index indicates that it takes, as a national average, 50% more to purchase a Big Mac in Brazil vs the US, while the Gini Coefficient, if I understood it correctly, says that there is a greater disparity between the wealthy and the poor in Brazil than there is in the US. Combined, it seems to indicate indeed that life is harder in Brazil for the average "middle class" person than those of us in the US appreciate.

Cheers and Regards

Edited April 14, 2012 by bphlpt

[dencorso]

We also use the 40h week, which means 8h of work per day. But the minimum wage is the pay for one month of work. As one year has 52 weeks, that creates the 4.5 weeks per month magic number used here in many calculations regarding work. However, people here receive per 13 months, the 13th being paid 2/3 in Nov, then 1/3 in Dec, and being less taxed than other month's wages (for those who are taxed). People on Minimum Wage are exempt from income tax, but still have to pay 8% of their income to a compulsory state managed fund (to which the employer contributes further 12%) part of which, can be used on the event of loosing the job to get by until finding another one, and part of which fuels the National Retirement pension fund.

So, by using 4.5 weeks per month, 40h per week of work and US$7.25 per hour, I'd conclude (using the Brazilian magic number for the weeks/month) that the Minimum Wage in the US is US$1305.00. And, IINM, there are just twelve payments per year in the US, so that would mean US$15660.00 per year (Which compares well with the more realistic 52.14 weeks per year,40h per week of work and US$7.25 per hour, which gives US$15120.60) .

Now, with Brazilian 13 months and Sao Paulo's US$370.00 per month, we'd have US$4810.00 per year. Then the buts begin: but the Big Mac Index (by applying it, the US$4810.00 is reduced by 35% to account to the higher costs here, becoming US$3126.50), but the Gini Coefficient and but the unemployment rate (6.0% in 2011). And Brazil has about 2/3 of the US population.

The Big Mac Index indicates that it takes, as a national average, 50% more to purchase a Big Mac in Brazil vs the US, while the Gini Coefficient, if I understood it correctly, says that there is a greater disparity between the wealthy and the poor in Brazil than there is in the US.

Those conclusions are correct.

Combined, it seems to indicate indeed that life is harder in Brazil for the average "middle class" person than those of us in the US appreciate.

Yes, harder the average "middle class", much harder for the average "lower class", dire for the "lumpen". But, taken together, those comprise at least 95% of the population, if not more.

[bphlpt]

I'm not arguing your conclusions in the slightest, I'm just having a little trouble with the math.

At US minimum wage of $7.25 /hr x 40hrs/week x 52 weeks/year = US$15,080.00 per year gross income, assuming no vacation. I'm not sure where you got "US$3132.00 per year". If the Brazilian yearly income at minimum wage is really just "US$4810.00 per year", then the comparative situation is even worse that you were portraying it, especially after you factor in the Big Mac Index and the Gini Coefficient.

Cheers and Regards

Edited April 14, 2012 by bphlpt

[dencorso]

I'm not arguing your conclusions in the slightest, I'm just having a little trouble with the math.

I had made a quite botched calculation, which now I've fixed, thanks to you.

If the Brazilian yearly income at minimum wage is really just "US$4810.00 per year", then the comparative situation is even worse that you were portraying it, especially after you factor in the Big Mac Index and the Gini Coefficient.

It is. Yes.

So, to get back to the start, people earn 1/5 and the hardware costs twice to thrice.

[dencorso]

Here are two real life examples:

San Disk 32GB Extreme Pro 45MB/s SDHC Card (at memorytec, Brazil) R$427,49 == US$266.19 (at amazon.com: US$89.99)

USB 3.0 Transcend PDU3 PCI-e Add-On Board (at memorytec, Brazil) R$56,99 == US$30.15 (at amazon.com: US$17.99)

In this context, it's no wonder people takes much longer to dump an older machine for a brand new one, and that upgrading existing boxes and repairing hardware (like changing for tantalum all caps in a graphics card with some bust caps) makes sense.

[os2fan2]

Windows 3.1 came out in Easter 1992. I got a copy when i bought my first computer at that time. It was a 486 DX 33 with 8 MB of ram, and 120 MB of fixed disk, a 5.5 floppy and a 3.5 floppy, and vga graphics. Except for the graphics, it was a pretty mean machine in its day. The usual fare was a 386 with 4 MB and 40 MB disks were just going out of supply, but were still being sold.

When you look at it, it looks pretty much like Windows 3.0, except a lot less grey. The shell program is 'progman' , which was considerably less potent than say, the DOS menu the computer came with, or even the program manager in DOSSHELL. You can't put groups inside groups. You look over the shell, and see what's missing.

I really did not get use to the shell. Shell replacements were one of the largest classes of Windows 3.x apps. I think i tried 15 or so. TakeCommand, Praxim, WPSShell, Norton Desktop Windows, Bubba, Command Post, Becker tools shell, MS-DOS Executive, Spider, even Solitaire.

The usual sport was 'program thinning'. DOS and Windows took up something like 20 MB, but you could thin out the rubbish to 10 MB or so. Norton Desktop (8 disks = 20 MB), and Becker tools (2 disks = 4 MB), were thinned down to 1 disk 750 KB. Five versions of Civ1 was thinned from 10 disks 15 MB to 3 disks 4 MB. Even DOS was culled.

After 4DOS, Rexx, one sinks into OS/2 and multi boot, and then multi-multi-boot, and then even using one copy of windows in multi-boot with network-like install. OS/2 is a wonderful system to learn about Windows, and Windows is a wonderful system to learn about OS/2. About themselves, not so good.

I rewrote the Windows install about 86 times, i suppose. I even replaced the tutorial with something useful (a welcome help system). It is still being developed. You could fix up most of the bugs in Windows by using the Win-OS/2 files, but i wasn't all that game in replacing things like winoldap.mod, etc. But the rest works.

[Fredledingue]

CoffeFiend,

What I mean is the phenomenon (rather than the thechnical evaluation) of poeple facing something new, which looks great and promising but without real application.

Other important innovations like SATA, USB2, etc had immediate use.

Why w3.1 if it's to launch only DOS softwares?

Why x64 if it's to launch only 32 bit softwares?

The move forward took much longer than merely a new piece of hardware which comes with a driver.

In these two cases, the entire software and hardware industry had to update their products to make the innovation fruitfull.

The concept of windows was not a hardware advance, but it forced the software industry to completely rethink the way interfaces are designed and adapt to a wide range of new possibilities, notably the use of a mouse.

While not new when w3.11 was released, it was new for MS-DOS users, most "home users" at the time, GUI enabled computers were still rare high-end machines.

At least for the PC world, there was a huge albeit gradual shift from DOS-launched, often mouseless and 2D apps to 3D windows apps.

The x64 was not as visualy spectacular because an x64 PC looks like a 32bit one, but it was the first time in almost 10 years MS made a core change in the OS, along with a core change in the hardware (processor).

32 bit was so much the norm that you would be yawning at the mention of it. Today it's more exciting.

Interresting link: A Short History of the GUI and the Microsoft vs Apple Debate

_________________________________

[os2fan2]

One must remember that DOS programs were designed to run on stand-alone machines, one at a time. Windows, and pretty much any multitasker, has to deal with shared use, such as spooling printers, file resources and other things which a computer needs to share.

It's interesting to note the only way i could get DOS games (like DOOM) to run on some boxen, was to run it in a winoldap session: a DOS box under Windows. This is because Windows can provide more memory than raw DOS could.

Likewise, in restricted memory systems, one might run windows games under DOS, by setting eg 'shell=spider.exe' in a copy of Windows 3.1,

[os2fan2]

When you run something like DOS under WINOLDAP (win 3.1/9x) or DOSKRNL (OS/2) or DOSVM (WinNT), the DOS code does not run in the VM.

Run for example, MSD under Windows NT or OS/2, and you will see that things like MSCDEX and Mouse take up a lot less code than under native DOS. This is because the calls are reflected out of the VM down to the host. Even functions of DOS could be reflected down to the kernel. OS/2's DOS emulation gives something like 630 KB of ram, with DOS loaded low, because the DOS calls are not dealt with in the VM. The 10 KB is the DOS data. All dos needs to do is update the data tables, not be present itself.

Likewise, a newer OS can introduce things that the old OS does not know about. Running things like WP and Lotus 123 in Windows VM's allows the prints to be spooled even if they start printing together. Running 32-bit proggies in 64 bits might allow each 32-bit system to run in its own space, even if the underlying 32-bit OS is shared. In other words, you run the base 32-bit OS in one VM as a DLL, and then run each application in its own VM, with a very tiny bit calling the master VM in the way that proggies in the main system calls DLLs.

[dencorso]

You could fix up most of the bugs in Windows by using the Win-OS/2 files, but i wasn't all that game in replacing things like winoldap.mod, etc.

Would you please kindly elaborate? I hadn't heard of using OS/2 files in Win up to now.

[Fredledingue]

Yes, finding new updates off non-MS OSes, would be awesome.

But is it realistic beyond a few DOS tweaks?

______________________________

Edited April 17, 2012 by Fredledingue

[os2fan2]

You could fix up most of the bugs in Windows by using the Win-OS/2 files, but i wasn't all that game in replacing things like winoldap.mod, etc.

Would you please kindly elaborate? I hadn't heard of using OS/2 files in Win up to now.

Most of what's in \OS2\MDOS\WINOS2 is an IBM compile of Microsoft's Windows 3.10 code. It has a handful of DOS files in it, and some icon changes (including winoldap).

It's pretty easy to work out the hacked files, since OS/2 for Windows adds such files to a Windows 3.1 setup. These have different names, and use different functions to the standard DOS features. There are proggies that can add files from MS-Windows to Win-OS/2 to have the same effect.

Some of the files, like write.exe are identical. Others like progman and calc.exe, are different (the OS/2 version does not have the bugs). Winfile works quite well under both, i suppose. The help about dialog lives in 'shell.dll', so when you call this up, you get 'win-os/2 3.11' under my Windoze soup.

After 1994/01/01 the IBM/Msft joint agreement went into another stage, and Wfw 3.11 was brought out with some modification that adds incompatibilities to OS/2 for Windows. IBM did fix this up, in a later release. In any case, you can run Wfw 3.11 under standard mode, (since wfw 3.10 does), but you have to lift some files from win 3.11 or 3.10 to do it. On the other hand, setup in wfw 3.11 does not handle standard mode things, even though space exists for it. You have to manage these adjustments manually.

Still, nearly all of the utilities from OS/2 i have used under DOS, Winfile and Printman comes from Wfw and other files lifted elsewhere. If you put shell.dll and the two ole*.dll files, Windows will tell you that it's win-os2.

Note in the picture it says 'win-os2 version 3.11'. It's running under DOS, there never was a win-os2 3.11.

Edited April 18, 2012 by os2fan2

[jds]

This is because Windows can provide more memory than raw DOS could.

Not true! Back in the day, I had a very memory-demanding application, whose needs could only be met in "raw DOS" as you call it, not in Windows.

What you say may apply only if you compare to an un-optimized DOS system, or perhaps one optimized with the 'MemMaker' utility. However if you are skilled, it is possible to achieve up to 120K more available memory than what 'MemMaker' can achieve. With DR-DOS, even more is possible, for example, I had a CD-ROM driver, mouse driver, LAN card driver and the bloated MS network client drivers [edit: and a disk cache] all loaded, and still almost 640K of available conventional memory.

Joe.

Edited April 19, 2012 by jds