In the 90’s I worked for a large retailer. Our centralized systems were on IBM mainframes, with some Unix servers for specialized work. We ran a small Unix server - really just a decent PC - in the stores. We had very low end machines running DOS-based terminal emulators for POS, etc.
Senior management wanted a some kind of communication platform for store employees. Something that could handle shared, broadcast comms to all employees, but also individualized comms to store managers, and ideally even questionnaire-like forms for surveys and the like. The entire system had to be secure and tightly controlled, e.g. employee A at store 1 shouldn’t be able to email employee B at store 2.
At some point we’d adopted using Lotus Notes at the HQ as platform de jour. (previously we were all in on PowerBuilder apps, remember that?).
It was no surprise, then, when IBM flew a one of our VPs to Cambridge (or maybe Boca Raton as part of another deal…) and he came back and announced we had purchased something like 50,000 seats for Lotus Notes and were implementing it as a store comm system.
A colleague and I were part of the architectural team. We didn’t have much of a formalized practice in those days, but we’d both been involved in the design and implementation of a lot of key store systems, and we weren’t idiots. We pointed out we’d never tested this sort of thing.
The Lotus team suggested installing Notes in a test store. We pointed out we had no compatible Windows equipment in the store. They suggested accessing Notes remotely. I can’t recall how. Unfortunately, our store networking system was Very Small Aperture Terminal satellite-based. High bandwidth, but also high latency, something like 3-4 seconds. Interactive apps over VSAT just didn’t cut it.
Now my memory of all details is a little fuzzy after 30 years, but as I recollect, Notes could effectively work offline, and then synchronize when back online. This was a cool feature for distributing apps and data. So, of course, the Lotus teams said the obvious, “Just purchase a separate server for Notes in the stores, and we can sync to that, and then all access will be local.”
Did I mention this was Retail?
Suffice it to say when we told Sr Mgmt the solution was going to require purchasing a dedicated server for every location (we were growing like crazy those days), the whole thing came to a screeching halt, and the 50K seat deal was cancelled. Someone lost a nice commission.
At the time I didn’t blame IBM. They were just acting on very incomplete information. Our VP, and old MF guy who now had responsibility for “Distributed Systems” as well, should have had an internal review before committing. I suspect he didn’t get any IBM Christmas cards that year.
Did we adopt standard email for comms? Oh, heck no. We rolled our own Frankenstein implementation of a backend and client, replete with our first C app on MVS. That’s a story for another day.
Notes stayed alive as our internal email, etc client for quite a while longer. I even got it running under Red Hat Linux, courtesy of Wine. Eventually the wheel turned, and we moved to something else. Was it better? No.
An absolute classic of "we've done no requirements analysis at all, but our exec went drinking with the wrong people and that now defines the solution, regardless of the problem".
In a former life about ten years ago I transitioned to managing a long-established, mature development team. Some of the developers had a couple of monitors. I think the largest was maybe a 19" model. Our business users, who sat on the opposite side of the building, and who they visited frequently, sported 3 monitor setups, each 24", maybe even 27". The main point was that our business users had more and substantially better equipment than the dev team. The devs were not happy about this. I was not happy about this.
Thankfully I had a reasonable boss. Apparently the argument, "we can't design UIs, we can recreate how our users actually use our product" hadn't been tried before or at least hadn't been successful. Suffice it to say I put in an order that recreated the UX of our business users.
I shared this news with my peer managers in the department. They followed my tack, and soon enough at least 3 other teams were rocking decent equipment.
Then the VP got wind of it, and said, "No more orders for new equipment until the existing equipment is EOL", or something like that. You know, okay, I get it, we're talking a lot of monitors. But, dammit, people were just a little happier, or maybe just not as pissed off at one thing we could control.
I didn't even try to expense the 43" 4K TV I used for a monitor at home.
Every time I’ve worked at a company where teams started picking their own hardware and expensing it, it ended like this. It usually started with reasonable orders, but once some teams notice that expensing hardware is working they try to push the limits of what they can get away with.
Ended up with weird situations like software and accounting people rocking those ultra expensive Apple cinema displays while sitting in their new Aeron chairs at their motorized standing desks which were always in the sitting down position.
Then someone gets shocked by the office expense line item at quarterly review time and cuts it off for everyone to try to balance out those expenditures for the rest of the year.
So we’d have a few people in the office with tricked out workstations and high end gear and then everyone else was on the same generic monitor and chair that was decided for the whole company long ago.
Even the most generous company I worked for that made a point of supporting engineer’s workstations eventually had to cut it off and standardize because the purchases were getting unjustifiable. There was also a lot of stuff like nice monitors or computer upgrades “disappearing” where you couldn’t really pin blame on anyone, because it was actually common for people to swap hardware around.
The only sustainable solution I’ve seen is when engineering gets together and argues for a reasonable set of computer gear and also a phased upgrade plan that doesn’t involve one giant expenditure in a single quarter.
Yeah we had the same thing. We had one person choose a massive curved ultra wide screen and then decide they didn’t like it so it sat in the hall beside their office for 6 months.
> the only sustainable solution I’ve seen is when engineering gets together and argues for a reasonable set of computer gear and also a phased upgrade plan
Design by committee doesn’t work IMO, and the phased upgrade plan works until bob decides that Fred got a new PC and wants the new GPU or feels stung by the fact that he got an older model by virtue of being upgraded 6 months earlier.
The only way I’ve seen it work is when engineering accept that they don’t need the absolute latest and greatest, and the directors plus agreeing on rolling stock upgrades. On my last team, I had someone argue with me that they needed a GPU upgrade from previous gen flagship to current gen flagship, when our game didn’t even require the tier below their current card to develop. They had the newest workstation of any team member at the time.
I think that providing modern monitors that make sense is probably the easiest way to please employees. No matter how excellent the salary is, if they are stuck working on old, small, blurry, faded screens, they'll be unhappy. It also gives a bad impression of the company as a whole. I did L1 tech support a long time ago, and people were overwhelmingly happy to see me coming to replace their old CRTs with LCDs.
I saw Sonny Rollins at the Montreal Jazz Festival in 2005. I had written a paper that was accepted at a conference, and thing realized the festival was going on at the same time. Unfortunately I learned this very late. Pat Metheny was the featured artist that year, and I'm a big fan, but all the tickets for him were sold out. I checked the lineup and realized Rollins was playing and snagged a couple of tickets for him instead. Honestly, I admit I mainly went to see him because I wasn't sure how much longer he would be playing.
I especially remember that he played "A Nightingale Sang in Berkeley Square" - my wife and I still talk about it.
I was fortunate to go to a high school that acquired a single Apple ][+ kept on a rolling cart that was locked in the chemistry lab closet for safekeeping. Every school day at 2:15PM there was a race to see which student could claim it until they finally kicked everyone out of the building. Thankfully I had pretty fast legs and a nearby last classroom. Our first “instructor” had the good sense to know he was never going to keep up with a bunch of rabid students; I don’t recall any of the lessons, but I recall the experiments we devised.
In 1983 I purchased my first personal computer, an Apple //e. By then we had an entire lab of Apples and Franklins, but I no longer needed to stay. The setup at home was more convenient, but the limitations imposed by the previous setup had a powerful focusing effect: hand-written programs, carefully reviewed and mentally simulated.
I had a similar situation, first computer access was a single Apple (Packard-Bell) ][+ on a rolling cart that normally lived in the closet of the 6–8 grade science teacher at my elementary school. I would stay after school every day to write programs on that, first in BASIC and later in 6502 Assembly. Floppy disks were a luxury to me, let alone actually owning a computer, and all my programs were handwritten in spiral-bound graph-paper notebooks. Even though I wouldn’t own my first Apple computer until 15 years later, I was still a life-long Apple fan, even when I was stuck on other operating systems over the decades.
To this day, my mental model of a computer is still a 64K Apple ][+ and my imaginings of how I would extend it given the capability to do so. My idea of a graphics system that didn’t rely on main memory to represent the bit map turned out to be predictive of how graphics cards now work, albeit constrained by the imagination of a high school student in the 80s.
My wife and I visited Reina Sofia last year and saw “Guernica”. That entire wing is fantastic: the section leading into the room covers war posters and other contemporaneous works, leading you into a room dominated by the painting, with side galleries focusing on smaller studies and post-work of characters.
Aside from the subject matter I was not prepared for the size of the work. It is one thing to see on a monitor or handheld screen, and quite another to see it full-scale. It is a qualitative change, not merely a quantitative one. So while this high fidelity picture allows one to study technique, color, and changes he made while painting, I think it misses the point. I say this without having read much commentary or critique of the work, but I imagine Picasso was so overwhelmed by Guernica that he wanted any viewer to overwhelmed, too. So if they do move it, I hope it is in a similar setting - in a moderately sized room, on a wall not much larger than the work itself, inescapable.
As ridiculous as it may sound, if you plan on visiting, plan two outings: one for Guernica, and one for the rest of the museum. That room and work are emotionally exhausting, and at least for my wife and I the intensity of that wing required a cooldown period.
Darn it. I literally just passed through and transferred trains in Maastricht in the last 2 weeks. What with the “anytime today” flexibility of Dutch and Belgian train travel, I would have loved to make a visit to the church a side quest.
I’m wondering who else recalls Charles Simonyi’s Intentional Software efforts at similar concepts? Not completely novel even then (1995) but mature for the time.
I have an Amateur Extra license in the US. I am also a Volunteer Examiner with a couple of organizations. Locally I am a Net Controller for a VHF Net. I am working on my ARES Level 1 Qualification.
My read on the reception of Steve’s post is that there are largely 2 camps, one of which thinks he’s given them a concrete tool to use, and the other of which thinks he has given them something to think about. I read his experiment as suggesting an agent architecture akin to Erlang supervisor trees, i.e. agents are cattle, not pets, and should be monitored and processed as such, with the obvious caveat that context matters.
Senior management wanted a some kind of communication platform for store employees. Something that could handle shared, broadcast comms to all employees, but also individualized comms to store managers, and ideally even questionnaire-like forms for surveys and the like. The entire system had to be secure and tightly controlled, e.g. employee A at store 1 shouldn’t be able to email employee B at store 2.
At some point we’d adopted using Lotus Notes at the HQ as platform de jour. (previously we were all in on PowerBuilder apps, remember that?).
It was no surprise, then, when IBM flew a one of our VPs to Cambridge (or maybe Boca Raton as part of another deal…) and he came back and announced we had purchased something like 50,000 seats for Lotus Notes and were implementing it as a store comm system.
A colleague and I were part of the architectural team. We didn’t have much of a formalized practice in those days, but we’d both been involved in the design and implementation of a lot of key store systems, and we weren’t idiots. We pointed out we’d never tested this sort of thing.
The Lotus team suggested installing Notes in a test store. We pointed out we had no compatible Windows equipment in the store. They suggested accessing Notes remotely. I can’t recall how. Unfortunately, our store networking system was Very Small Aperture Terminal satellite-based. High bandwidth, but also high latency, something like 3-4 seconds. Interactive apps over VSAT just didn’t cut it.
Now my memory of all details is a little fuzzy after 30 years, but as I recollect, Notes could effectively work offline, and then synchronize when back online. This was a cool feature for distributing apps and data. So, of course, the Lotus teams said the obvious, “Just purchase a separate server for Notes in the stores, and we can sync to that, and then all access will be local.”
Did I mention this was Retail?
Suffice it to say when we told Sr Mgmt the solution was going to require purchasing a dedicated server for every location (we were growing like crazy those days), the whole thing came to a screeching halt, and the 50K seat deal was cancelled. Someone lost a nice commission.
At the time I didn’t blame IBM. They were just acting on very incomplete information. Our VP, and old MF guy who now had responsibility for “Distributed Systems” as well, should have had an internal review before committing. I suspect he didn’t get any IBM Christmas cards that year.
Did we adopt standard email for comms? Oh, heck no. We rolled our own Frankenstein implementation of a backend and client, replete with our first C app on MVS. That’s a story for another day.
Notes stayed alive as our internal email, etc client for quite a while longer. I even got it running under Red Hat Linux, courtesy of Wine. Eventually the wheel turned, and we moved to something else. Was it better? No.
reply