Month: November 2017

Part III – The Solution

By Chuck Ritley & Ken Ritley

---

 

KEN:  So how did your Indian distributors  get into the picture?  Did you take the initiative?

I wish I was smart enough to take credit.  Coincidentally, one of the Indian principals had some business in the US with another supplier, stopped by for a get-acquainted meeting, and I had a couple of days to fine-tune some technical issues.  The topic of contract work came up.  “Would we be interested in contracting out program development work?”  Their technical staff were well educated, I had seen some of their product, and they could offer hourly rates far lower than the US.  We didn’t conclude anything then, but it gave me some ideas for the future.  These folks were a known quantity, and sold tons of our equipment in India with no complaints.  Clearly they knew what they were doing.

Let’s examine our coding deadline.   We had 8 programmers, and needed 16.  And IBM wasn’t slowing down.  We talked with the CEO about it.  We had other English-speaking assets, but India was the only one with enough excess manpower that was able to sell services.  So it was time to talk with them.  I had hoped for a trip to India, but we settled for the phone.

When we explained the scope of the project, the Managing Director in India showed an excellent grasp of the concept.  And offered us something we had not encountered:  a turnkey coding solution.  In the past, we paid contract programmers on an hourly basis, and had to constantly ride them to keep up production.  The Indians offered a hands-off deal to us:  they would provide up to 12 programmers, bring them to us, pay for their meals and lodging,  they would work under our supervision, and all for a package price.  If production goals slipped, they would bring in more help at no extra charge.

KEN:  Sounds like a good deal.  Almost too good, wouldn’t you say?

I was familiar with the “mythical man month”, and knew there was a limit to the number of people we could coordinate.  But the fact that they would take the risk was impressive.  And the total cost of the package was far less than I would have paid local programmers – if I could have acquired them quickly, which was no option in Silicon Valley at that time. Our CEO went off with his finance people and VP’s, and said “let’s do it”.  The bottom line was:  if we couldn’t get this done on time, we’d lose the game anyway.

KEN:  Today, of course, we bundle up projects and send them off.  How did you start an on-site deal like this?

Over the next two weeks, I renovated an old classroom with work tables, cabling, and 12 new terminals tied into a dedicated mini-computer.  I rented two small one-bedroom apartments close to the office, choosing “someplace that doesn’t have all that zoning stuff.”  The local government had some occupancy rules and I was cheating a bit.

The first to arrive was a guy who would be my main contact – let’s call him Krishna.  Well dressed, with excellent university English, Krishna explained that he would be in charge.  He rattled off some impressive IT credentials, and said that if I explained what I needed, then he’d interpret, delegate, and see it got done.  All I had to do was keep testing the coded output.  He checked out my workspace and said it was fine, and we went over to look at the 2 apartments.  I thought they were small for 12 guys but he said they’d be fine and I gave him the keys.

The following Monday, they all arrived — 12 programmers as specified.  I don’t recall their names as it’s been too many years.  I recall Kumar and Rakesh, but that’s all.  In my defense, the reason is simple:  only half of them were fluent in English.

KEN:  Now you have your team in place.  How did you get off ground zero?

I had a panic attack at this point, since our proprietary operating system and programming language were like the old Pick language – English-based.  Krishna calmed me down and explained that all of these guys – even those with no English – had been writing programs for their dealership for years and were proficient.  He said that the 6 who spoke no English were from Bangladesh, but that the Indians were all bi-lingual.   And they needed separation from the Bangladesh guys, hence the two apartments.

KEN:  Now you have your team in place.  How did you get the project started?

I got everyone into the classroom, where I had diagrams on the board and a projector with graphics.  (This pre-dated PowerPoint.)   I walked them through the logic of the system, slowly to let Krishna fill in those with no English.  Our own 8 guys, with their own modules to work on, sat in.  But at this juncture, all I could see was disaster.  Remember, this my first experience with an off-shore team.

We had a company van, and the 12 were hauled over to the apartments to get settled in while Krishna and I began working out a revised production plan. (I already had one, but it was geared up to the 8 in-house programmers.)  When I asked him about food, meals, and other necessities, he assured me that he would take care of all that.   My job was to assign tasks and test the output, Krishna would handle the work.   Bear in mind that I had been coding, and supervising coders, for years, but this was a new ball game.

KEN:  So give us a play-by-play of this new “ball game”.

So it began.  The next day they started coding.  Krishna must have re-briefed them that evening, and everyone seemed to know what to do.  They worked hard, with a work ethic I hadn’t seen before. No one broke for coffee, no one chatted with his neighbors, and no one wasted time.  There were problems, of course.  I made them print hard copies of everything so I could inspect all code, since no code works the first time.  And trying to keep up with 20 programmers was a strain on me, since all of the pieces had to fit together.

KEN:  So – so far so good?

All was not well at first.

Even with 12 extra programmers, we were falling behind our CEO’s release schedule.  Krishna was a task-master, drove the guys harshly, and they put in 12 hour days.  They kept to themselves, focused on work, and didn’t socialize with my guys.

A few weeks into the project, he asked if we had spare terminals.  We pulled 6 terminals and modems from inventory, took them to the apartments, and got them hooked into our development network.  Krishna planned to have them put in extra time in the late evenings on-line.

KEN: It sounds, at least, like the work was back on your schedule.

It was. But, at this point, it occurred to me – with much dismay — that I could somehow be stuck in the middle of a slave labor operation.

We were paying a flat fee for 12 people, jammed into two tiny one-bedroom apartments. (I was never allowed inside.) They fed themselves, so presumably they shopped and cooked.  Spare time was spent in extra work on-line. Saturdays and Sundays were just 2 more 12-hour work days.  I had my own terminal at home and I could see new code appearing on week-ends.

So my conscience prickled me – but not enough so that I was willing to tell the CEO that we couldn’t get it done.  Personal and business survival has a moral price sometimes.

KEN: So setting aside your Silicon Valley attitudes, how was it progressing?

 Now, honestly, the work wasn’t the highest quality.  The 12 turned out what 8 of my own programmers would have done.  I spent many hours going over code, editing subroutines, instructing Krishna on what was wrong and how to fix it.  After time, I had some rapport with the 6 English-speakers and instructed them directly.  Or I would have Kumar or Rakesh talk to the Bangladesh guys. That annoyed Krishna, but at this point I had a schedule to meet, and cultural customs be damned.  This, of course, caused some tension.  But, again, this was our first venture into new territory.

Weeks went by, but we made progress.

KEN: Progress in that you were keeping to your schedule.  But, how about the end result?

That was the over-riding potential hazard:  I knew from experience that when we went live, some code wouldn’t work. Does it ever? Programmers know what I mean.  So when the 12 went home, what we had was what we had.  And leave they did, with no fanfare, no long good-byes. I just handed Krishna a check.

As with any new system, there were bugs to be worked out over time.  But the code was plain, and well documented, so out own guys could handle that. And we got through the those bugs, and actually met the CEO’s deadline with a working system

KEN:  So in the end, everyone was happy about it.  I mean – happy with the results.

Yes, but it was a life lesson for me.  I was used to hard work and long hours.  Anyone who programs is.  But we never worked as hard as that Indian crew.   When I reflect back on it, I suppose it was a preview of things to come as more and more companies tried it.

From everything I could see, this was a win/win situation, or the guys wouldn’t have come and worked so hard.  But it’s hard to forget the 12 guys jammed into 2 rooms, working nearly 24/7.

 

Part II – The Challenge

By Chuck Ritley & Ken Ritley

---

KEN: Of course, today the ERP concept is an old one – it’s the bread and butter of companies like SAP, and you can get degrees at colleges that teach this stuff.  Now let’s get to your challenge specifically – what was your pain point?

I was supervising the coding of a large enterprise control software package. Somewhat later the industry term became known as ERP – Enterprise Resource Planning. We competed with IBM head-on, a tough crowd. This concept included all the functions a company needed from materials to accounting to shipment.  A few dealers who had made this their specialty sold the software, and so there had to be some margin for customization.

Our pain point – time, time, time. We knew IBM was almost ready for market. The design had many new features, and the market was much different than today. The first company to break into a market could easily wind up owning it.  So our management needed to finish our product and start selling it before IBM or we’d be playing catch-up.  And bear in mind, we were a mini-computer company, competing against the giant of the mainframes.

KEN: Today there are many possibilities to solve the time-pressure problem. Hiring good IT people is not that big a challenge – and if you are really under time pressure, there is a huge market of freelancers and even companies willing to take over coding on a fixed price basis.  How was it different for you?

Our programming team consisted of 8 “programmers” – today known as developers or software engineers. Based on our workload, that wasn’t enough. Now bear this in mind: it was a proprietary world. COBOL was standard only for mainframes, so it you wanted a competent programmer, you needed to find one with expertise in another language and train him or her in your proprietary language. Freelancing was just beginning, and pickings were slim.

There weren’t any degrees in IT as yet, and most programmers got their start by attending schools from one of the manufacturers.  IBM and Honeywell, in those days, had large training programs.  Most programmers available for hire had spent time doing mainframe coding for customers of one of those two giants.  And they might not have coded in the areas we were developing.

As matter of fact, I was recruited to Silicon Valley because there had been a shortage of good programmers and I had both IBM and Honeywell experience. So there were good programmers available.  But not overnight.

KEN: It’s certainly much different today. So what made you think of recruiting an international workforce?

Frankly, the international market had puzzled me for some time.  We had distributors in Australia, New Zealand, Germany, Belgium, Great Britain, France, India, and Japan.   What amazed me was that everything we published by way of technical documentation was in English.  The OS, tied directly to the CPU, had the usual English-like commands: edit, erase, format, etc.   The coding language was similar to an abbreviated version of BASIC, rather than the assembler-style used by mainframes.

What impressed me was that each non-English speaking distributor had to have an English speaker on staff.   From time to time, each would send techs to the US, always someone fluent in English, and we would train them.  They would go home and translate everything from OS and programming manuals to the electronic logic diagrams into the native language.  No small task.

But, at the time, the US was the only source of minicomputers on the planet.  IBM didn’t compete at that level.  For example, in Australia and NZ, we were the largest computer supplier.  So distributors who took this route could open new markets at the turnkey level. To support this, we had a programmer and field engineer who worked odd hours to support the time differences and they could provide answers via telex.  (No internet as yet.)

I was most impressed by the staff of our Indian distributors, who were a large multi-product electronics house.  Because English was common, they created their own software with our programming language.  Not surprisingly, some utility software, such as database systems, had market value in the US.

 

Final coming next: Part III – The Solution

Part I – Setting the Stage

By Chuck Ritley & Ken Ritley

---

Anyone working in IT has heard the terms offshoring and nearshoring.  Two things about them usually come to mind.  First, that they are supposedly about transferring jobs to low wage countries. And second, they were supposedly made possible by recent advances in technology, as Thomas Friedman describes in his book The World is Flat.

These statements are not entirely correct.  In fact, IT offshoring began much earlier, even in the late 70’s and early 80’s, much before the Internet. And the motivation was less about saving money, but having access to top technical talent.

As an IT guy in the early 2000’s, Ken helped build large offshore and nearshore IT organizations in India and Eastern Europe. This is what we’ll call “second generation” IT offshoring – in which the work is made possible by the Internet and remote collaboration. But in the 1970’s Ken’s father Chuck was instrumental in setting up and directing some early “first generation” work – in which teams of foreign specialists were brought into the U.S.

What follows is a discussion between Ken and Chuck to tell the story of how first generation IT offshoring began, the strong impressions it left on the people involved with this work, and to highlight some differences and similarities between what happened then and what happened now.

KEN:  I recall it’s all about mini-computers, which don’t exist anymore today. Can you paint the scene?   What was the year, what was the industry like, and what were you and your company working on?

You have to understand the IT industry in the late 70’s and early 80’s. It was proprietary and minicomputer-based.  Each company had its own processor design and, therefore, a unique OS and language.  At the main-frame level, companies had their own mainframe programming staff who did custom programs.  But the minis were meant for the small and medium-sized business.  Since those companies could not afford a programming staff, or development time, the minicomputer manufacturers sold turnkey solutions.

The set-up involved a network of dealers, most of whom were specialty houses.  For example, in a large market, there would be a number of dealers.  One might specialize in finance, another in distribution, or wholesaling, or transportation, or retailing.  Most would start with a standard accounting suite of software and then tailor it to the needs of a specific industry.  For example, a trucking company would have different needs than a food wholesaler.

The main point is this:  it was a turnkey solution for the end-user, both HW and SW.

Now, to make this happen, the computer manufacturer would provide the basic accounting package: sales, orders, accounts receivable and payable, purchasing, inventory, general ledger, shipping, etc.  These are root functions that every company does – but a trucker does them differently than a clothing retailer.  They were designed to be customized.  Having customized one for a specific customer, the dealer now had some expertise and could sell the same customized version of the package to other prospects. The concept of having standard modules which could be tailored is probably the root of SAP and other systems like it.

KEN: That’s certainly much different than today.  Today the hardware is standard, the software and programming languages are standard, and the main differentiator between companies is the domain or solution.  Can you provide some details about your domain and solution? 

The system of customizing basic accounting packages was used by most of the mini-makers.  And all of them generally had the same set-up of specializing dealers.  But that wasn’t enough.

There was a need for more complex software suites to handle much more than basic accounting, things that dealers with limited staff couldn’t handle.  Manufacturers and distributors have complex operations and want to use computers for operations like scheduling work, planning lead times, and materials acquisition – far beyond the reach of normal accounting.  So we began to specialize in full-function enterprise resource planning and manufacturing resource planning software. In other words, handling all of the functions a company would need to control every function.  These were also meant to be customized easily at the dealer level, since individual needs vary.  We also had a few others, but these were the most complex, and beyond the development scope of any of our dealers.

We designed it, trained the dealer, he sold it, and then tailored it to the customers needs.  Now,  here is where the problems lay: creating these complex software modules required not only excellence in coding, but expertise in the mechanics of a business.  The solution was usually to have a business analyst working with the programmers to spell out the flow of work in terms that they could turn into code.

 

Coming next: Part II – The Challenge