Lotus Notes and the dangers of starting from scratch(buttondown.com) |
Lotus Notes and the dangers of starting from scratch(buttondown.com) |
Yes, the UI was clunky, but it was quick. It was McMaster-Carr quick. The mail client was never a barrier between you and your emails. These days I miss Outlook, as the new web-based solution we replaced it with is even worse.
My reasoning (which I ranted on extensively for the rest of the interview) was that mandating Lotus Notes as a tool internally, let alone continuing to sell it, reflected very poorly on how IBM felt about their employees, customers, and products. Examples:
* for YEARS IBM had to ship a "killnotes.exe" utility to their own employees because they couldn't figure out how to gracefully exit the app - and that's just one papercut out of hundreds;
* they misrepresented the popularity of the product, as the number of seats sold was tied to other bundled deals - most customers wouldn't touch the product;
* they continually gaslit employees and customers that Notes was great for common workflows (email in particular, searching in general) where in fact it was overwhelmingly a failure - and anyone who complained "didn't understand the product";
* they would send out multiple-choice surveys on "How would you like to improve Notes?" with every question and possible answer tied to developer-focused and database features with nothing allowing feedback on the primary use-cases that were driving users up the wall
In the end IBM's stance on Notes was either tone-deaf, incompetent, mendacious, malicious, opportunistic, extractive, callous, or shameless with no sense of pride in product.
I'm still bitter. Man, I hated that product.
a. Everything you think is new was created in the 1960s by Don
Bitzer @ UIUC or Doug Engelbart @ SRI. (and probably also
Robert Creasy @ IBM.)
b. Lotus Notes (and Groove) were products, not protocols. If
you wanted to use them, you used Windows and the server and
client from the vendor.
c. Lotus Notes only sort of worked over the intarwebs. It had
serious problems traversing firewalls, and technically worked
over VPNs, but was nearly unusable as it required you to
either replicate your AD server or use AD over the VPN (I do
not recommend the latter option.)
d. Like JIRA, in order to do anything interesting with Lotus
Notes, you effectively had to have admin privileges. This
isn't strictly true, but it's effectively true. Both JIRA
and Lotus Notes were designed for organizations whose thinky
bits were managed centrally. Have a site or department
specific process? Better schedule time with your admin to
talk about how they're going to model your work-flow because
there's no way you're going to be able to build it yourself.
Groove was a little better, but it was disturbingly easy for
the brittle XML based DTDs used to model data to become out
of sync between users. Guess it's time to restore from
backup.
e. As much as I loved PLATO, it was a centralized / mainframe
type system. Of course it was easy to coordinate via a notes
file, it was a single file on a centrally shared,
non-replicated volume that (usually) had very loose write
permissions.
The reason notes (or it's bastard offspring Lotus Notes)
didn't take over the world was it put all the thinky bits in
one place, probably locked down by a BOFH. There's little we
learned from the micro-computer revolution from the 70s and
80s, but what we did seem to get is we all hated being
dependent on that one guy from accounting to write a few
COBOL programs for us so we could see what our sales figures
looked like this month.
It seems weird we're comparing PLATO NOTES with Lotus Notes
email. PNOTES (Personal Notes) was the equivalent of email.
NOTES was more the equivalent of Usenet News.
And Notes was an application environment w/ data replication
that had a shim layer on top of it that looked like an email
system. It was more like Hypercard with git-ish semantics.
(Which now that I think about it, wouldn't be that bad of a
product.)
But... after more than 50 years in tech, I still use Thunderbird
as my email client. If I could figure out a way to get it to
turn off HTML mail I would. I'm on the vanguard of the orange
catholic butlerian jihad. Not only do I seek a prohibition of
thinking machines, I'm also quite suspcious of richly formatted
text.
Maybe the reason (Lotus) Notes failed was that people kept
thinking of it as an email client. Raymond Lowey talked about
the "Most Advanced Yet Acceptable" principle. Meaning you can
push product design only so far before people no longer
understand how they're expected to integrate it into their
lives. Lotus Notes was the poster child for this concept.I still remember the Notes/Domino Server TCP port number - 1352. I've used it so often in firewall configurations or communications with customers and firewall teams.
I had servers hosting data for replication to customers over the internet. (Including patent data, very sensitive in both security and timeliness). We never had any issues connecting servers over the internet. It was easy, reliable and simple.
It was also very secure due to the PKI infrastructure built into Notes, the ability to require encrypted network traffic, and the various layers of security on the server.
Plus the NRPC protocol has its own proxying feature, called "Passthru Servers". That allows clients or servers to connect to other servers via an intermediate server - and that intermediate can actually blacklist or whitelist other clients/servers from using it. It's all very easy to implement, a core part of the product that any Notes administrator should know about.
But what really confuses me is you talking about AD. You can do some integration with AD and Lotus Notes, but out of the box it uses its own directory and PKI infrastructure. That's famously one of the things people held against it - that it didn't use AD or NDS. If you did turn on the AD integration then I guess you'd need access to AD, but that's your decision - and the downsides of it are much the same for any local client software you're trying to authenticate via your AD.
I simply can't reconcile what you've said with my experience. So I'm very interested to understand what you're getting at there.
view -> message body as -> plain text :)
In v153.2.0ESR, it's More --> message body as --> plain text
and it absolutely does work.
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.
There's exactly two people who LOVE Lotus Notes. One is the boss, cos he's never wrong, and the other is the Lotus Notes developer/integrator, cos he's getting money.
EVERYONE else in the org goes "Uggghghghhh, not Lotus Notes!"
We had our "email blast" code in the founder of the companies mail database. When we launched new features or had various announcements, you would log on to Mark's email database and run those features in the drop down. Of course there were various permissions in check that only certain people could do (or view) certain things in the database, but that's where it ended up...
I always thought the data/design replication was cool as well. I had to bake in what I called "relational fields" into the documents to be able to perform some standard relational database type of work, but it was really easy to work with overall. Maybe too easy - as one person mentioned in the thread, there were hundreds of little custom apps built by anyone/everyone in the company, because it was just that easy.
Yes, it could do all sorts of things, but it never seemed to do any of them well, and it was a horrible email client. Perhaps it was revolutionary in the 80s, but by the 00s it was clunky, heavyweight and painful to use.
We faced a situation where we had about 4,000 custom Notes apps to migrate, a small team and an extremely limited time. (and varying levels of complexity in each app)
What we found we could do, was "expose" the apps as "Notes Web" apps, and extract much of the UI/UX, and other artifacts - then we wrote conversion utilities.
We were able to get the migration time for simple apps down to the time it took us to demonstrate our framework... (20 minutes per app).
However - quite honestly, the more I saw of Notes at the time, the more I asked "Why migrate?". As the feature-set of early "SharePoint" was not even "at par" with Notes - even now, with the myriad of M365/desktop/app integrations, it is generally "just ok" and does not feel "cohesive".
Notes and Sametime also worked very well on Linux on my IBM work laptop, supported by support, official Linux distro inside IBM. That was mid-late 2000's when many companies hadn't even thought of trying to support both Windows and Linux (or Mac).
I was briefly what is now called a developer-in-test on Sametime after Lotus Development bought them. I was untrained in C++ so my bug reports for that API were frequently closed with the comment "No".
As an end-user, I welcomed Notes demise.
And don't get me started on its god-awful local/offline sync behavior. Outlook ran circles around it. Hard to believe from todays perspective, sure, but yes, Lotus Notes was so much worse.
We were asked for something quick and easy, but decided to put the effort in to make it the correct way for the Exchange architecture.
The bulk of that code is still there 30 years later.
Sure I guess if all you cared about as an end-user was email and Windows-like UI, then I guess I understand. But you were missing a ton of other things.
I don't know if I want to fault Lotus for that. As someone who juggles between Windows, Mac and Linux on a daily basis, using wrong keyboard shortcuts is something that happens regularly. For most users who only use a single platform, that's even less of a problem -- you only need to learn it once.
This is how I feel about the article in general -- I am not convinced by its arguments. Many of the design decisions make sense to me even today, and are more of a necessary limitation than anything else. I don't see how "at most 8 windows" is any worse than Windows' 256 character limit -- 8 windows are plenty at the time, and you need to decide to stop optimizing for someone who regularly has 20 open windows for no reason and focus on something else. Handling more windows can come later.
This article just reads weird to me.
The paradigm of everything being a database with documents was ahead of its time, but the real power was in LotusScript. It was basically just Javascript and that allowed me to transition to a web dev role later on. I have mostly fond memories of Notes.
The modern iteration is Notion.
Having said that, once you're locked in (and I don't just mean email I mean the hundreds of custom applications your company has built) it's a slog to get off. Meanwhile you still pay the full licence costs whether you have one custom app or a hundred.
Same thing with Google Wave, really.
The first few years were a little heavy on RAM usage, so people thought it was slow. To give you some context, this was back when 16 MB (megabytes, not gigabytes) of RAM was a thing. But as RAM increased, Notes was no longer a problem.
For example, Lotus Notes 8, which I still have on my personal PC, uses about 11 MB just after starting. Then it adds roughly 10 MB for each "app" you open. Today, Outlook alone uses around 170 MB just after starting.
It was incredibly powerful. And migrating a Domino server was incredibly easy. Basically, you just copied the DATA folder. I remember doing this when we migrated from Windows 2000/2003 to AIX.
In the same way, the Notes client was basically a portable app. You could just copy the NOTES and DATA folder, edit notes.ini, and have it running.
So sad to see it go.
It is. And it's the main thing why my boss does not understand why rebuilding databases takes so much time. For him its just a little thing, some fields and off you go...
Unfortunately it is possible for databases to be copied, mixed and matched and deployed somewhere else resulting i a pile of duplicated databases without structure.
The issue with Lotus Notes is that there was too much low-level machinery exposed directly to the client, and things you would expect to work with a basic email client, like .ics invites, simply never works right. There was no good dev environment for it, and most of the "develop your own app" feature was locked by corporate policy anyways so we pretty much had the worst of both worlds.
All in all, not something I would recommend to actually use, the only thing it really helped me see is that OneNote's lineage from Lotus Note as I think a lot of the original Lotus Note devs jumped to Microsoft at the time.
It was one of AT&T’s attempts to “kill the internet”.
The people working on it were down the hall from me when I was at AT&T.
It only worked on Novell networking. I asked how they plan on dealing with various Novell limits and they’d respond with blank stares.
The plan was to restrict it to only work on AT&T networks. So, it was more of a marketing ploy than a technology.
When I’d ask “how could this little project beat the internet, which is huge and growing exponentially?” And was told basically they believed if they could act fast enough, nobody would know the internet existed.
And people wonder why I had no faith in AT&T management.
https://www.cnet.com/tech/services-and-software/at-t-dumps-n...
It was universally hated because it felt so clunky and old-fashioned, but it was being used for ERP-type applications with a lot of business process logic that had developed over time. I found it interesting that you could click through the interface and find a lot of capabilities that were vestigial, still present but unused because they were made obsolete by web apps.
I just searched and I see that it was rebranded HCL Notes in 2019. I don't remember that branding at all. It seems likely we were stuck on an obsolete, non-rebranded version, deployed on local servers.
I do have to say Microsoft has spent the last decade "upgrading" their product to be a similarly painful experience though.
Why Microsoft continues to burn goodwill for native Outlook to keep pushing webview "new Outlook" which isn't feature complete is beyond me. Unless we're getting webview Excel, there's no point because you won't be running Office on any other OS anyway.
Lotus Notes allows you to right-click on a "document" and see absolutely all data attached to it, directly in the UI, including access permissions. This can be immensely useful, because not all data could be part of its UI representation.
The closest to this today would be inspecting the JSON data on your browser networking tab to view what your app exchanges with the server, but its much more clunky.
We used it as a combo Helpdesk ticketting, email, wiki/doc tool and it worked so well. It took 20 years for other tools to catch up and webapps still arent as functional.
For what it was, yes.
For what a great many people used it for on the other hand…
It was a great platform for making document and workflow control applications, and was ahead of its time in terms of a proper PKI system for authentication/access/verification and access control lists and such. But many people hit a poorly made application and that coloured their opinion for the rest, and in the list of terrible applications I'm including the standard email support, which was a royal PITA, and most custom versions for that function. The general UX wasn't great either, not so bad that you couldn't put up with it for a great application but if you are hating the rest of the application then the general UX issues grind your gears too.
I have fond memories of filling in time sheets on Lotus Notes. Although it's always an annoying thing to have to do, at least the Notes form that the local IT guy made for it allowed for retroactive changes and comments. Never had any issues communicating with finance over discrepancies or billing. And it synced from basically everywhere, using a 9600bps mobile dial-up connection if needed.
You mark the database as read only, copy the .nsf file over, unmark the database as read only, done. Brightstor had a plugin for Notes that did that automatically and could do incremental backups with ease.
Move to Exchange and it's horrible single file mail storage solution which was inconceivable when hard drives weren't that big. 10,000 users? It's 100MB of mail so it can fit nicely in a single storage volume. Backups? Well, only at night because you have to unmount the file from Exchange, copy it over, and mount it again, which took a good couple of hours.
I miss Notes "simplicity".
We worked at Lotus Development, but we weren't from Iris or even in the USA. Notes was the golden goose & the risks from leaks were judged too high.
And once we had a design that the current userbase was used to, we couldn't change it easily even if the rest of the industry standardized on a different pattern.
But Lotus Notes's main strength--everything is a database, even email--was also its biggest UX weakness. The "email app" in Notes was built in the same low-code environment customers used to build custom apps. That was good because email got features like encryption and replication for free, but it was bad because we couldn't easily customize the UI to optimize email. Instead, we had to add general-purpose UI features to the platform and then use them in the email app. But that was more work and took longer. Outlook could improve faster.
I did a second stint at the same company in the teens and they had moved to outlook.
Users were trained in the new ways of email and some proprietary CRM app we developed, but IIRC at least half of the customers rolled back immediately because of user protest over user interface and missing/non-working functionality, one customer even went under.
In principle Lotus Notes was the best, but practically, it was horrible.
It was funny I recall never having much to do on the domino server, and even the client was pretty stable. When we migrated to Outlook there were non stop server side issues.
The fact it had a built in DB was really useful for extension as well.
Imo it was ahead of its time and heavily hamstring by Outlook dominance as any nee user landing hated it simply because it wasn't outlook
Yeah, Notes had its shared database features, and you could see remnants of IBMers using them at the time, but most such applications been ported to Java CRUD apps. Also Lotus Sametime (it's like Slack) became a much more important collaboration tool.
The same problems keep being solved, and then nation-states get involved, and eventually the technology shifts to something else because the game has played out as far as it can in terms of those nation-states. It is not really so much as a steady march of technology, as much as it is a quiet kind of informational arms race.
The movie studios were pretty sure that people would want to buy movies and watch them at home. What they weren't sure about was how to price the movies because this was a brand new market that had never existed before.
This led to the price for a movie on VHS/Beta ranging from $10 to over $100 because no one knew what the actual demand curve looked like.
I feel like the 1990s was a lot like this but with UI/UX. People had opinions on what consumers wanted but the market was expanding so quickly that no one knew for sure. Firms just tried things and, in hindsight, some firms seemed "smarter" because they picked the feature set that customers ended up liking.
Used well, Notes had tremendous power. Enough to make the deficiencies of late-'90s/early-2000s email tolerable.
Used badly, Notes could easily fuel the kind of rant that ended with you being angry the office windows didn't open, because the only sound that might calm your inner rage was your computer's second bounce off the pavement.
But that could describe a lot of IBM software at the time. Often good ideas, but rarely enough structure to protect people from inflicting their own silliness on themselves.
This article doesn't feel like it was written by someone who was using these platforms, nor like anyone was there at the time was consulted.
For example, focusing on F5 is odd. F5 is not the Office refresh key. F9 is. Excel uses it to recalculate. Word uses it to refresh fields. Outlook should have used it to fetch/refresh mail, but for some reason they went with the Windows/IE platform teams' choice of F5. Microsoft did actually switch back to using F9 in Outlook, and there was a significant outcry from their users. None of this is mentioned.
Similarly there's a screenshot of a messagebox stating that there are too many windows. No context given, but I'd bet it's from Loutus Notes 4.6 or earlier, which used MDI. MDI is limited to 9 windows. I hit that limit in a lot of applications back in the late WIndows 3.1/early Windows 95 days, at the peak of MDI. Notes Release 5 moved away from MDI, and gave the user a tabbed interface. The Exchange Mail and Outlook clients never used MDI, opening a new window for each mail - which may or may not be a better solution, given that Windows didn't have task stacking on the taskbar at the time. Nine tabs in Notes was just one taskbar entry, nine windows with Outlook was heading towards a serious usability problem...
I'd like to note that the Notes client didn't benefit from the network effects that Outlook did. At the time many of the complaints in the article were being made software packaging and deployment was in its infancy. Many environments made Notes and Outlook/Office part of their base build making updates a difficult thing to accomplish. Not to mention that the installers were quite large, taxing many networks of the time. But Outlook had the benefit of being part of Office - sometimes people wanted the latest features in Word or Excel, so the effort was made to update the installation. Notes had no such advantage. A similar network effect arrives when it comes to licensing, given that larger organisations will have volume licensing deals.
By contrast on the server side Exchange Server was a complete PITA to upgrade. You don't merely install new software, you build an entire new infrastructure and migrate mailboxes over to it. Very expensive, especially in the days of physical hardware. Whereas Lotus Notes Server/Domino was very easy to upgrade or extend.
As a result many Notes installations were up to date on their backend but not the frontend, whereas many Outlook installations were the other way around.
Anyway, I've gone on for too long - much like all the products we're discussing here. ;-)
Some of the different decisions everybody made were for good reason. I know Apple and [later] Microsoft spent a large amount of money on human factors engineers doing real experiments to figure out what worked. I have never seen evidence that Lotus spent the money to see what was a good idea - they just made it work.
I was administering a small Domino server and using Notes for email. I think it was the worst thing I worked on at the time.
MSFT was so dominant back then that Gates had to rescue Apple to keep up the appearance of a fair market!
The web killed Notes, not Microsoft.
Also depends if your "platform" extends beyond just the OS. If platform includes IDE/editor/terminal, then a lot of users are juggling multiple keyboard schemes from different eras (ctrl-C has entered the chat).
This article is more about "first mover disadvantage" or "handicap of a head start" where an early implementation is leapfrogged by competitors that don't have legacy baggage.
In a way yes, IBM wanted notes to be an email clone. IBM took cc mail (I think) and shoehorned it into Lotus notes.
To me, that ruined Notes. If I get the versions numbers right, version 3 was great, IBM/Lotus came out with v4, which to me was a bit worse, and each new version was worse than the prior. Instead of focusing on emails, IBM should have focused on making notes servers easy to install and maintain and keep the application portion relevant.
By worse, it was the application portion that degraded. If you saw how cool the applications were with v3, you would wonder why Notes never became a "thing". Even third parties were creating applications for Notes at the time, some were very good. But to me IBM ruined that business by focusing on only Email part of Notes. Now AFAIK, there are no 3rd party applications for notes.
Maybe the company that bought notes from IBM is trying to get back to the roots, but I kind of doubt it.
Office 2007 introduced the ribbon and got under pressure by the web (GSuite).
But the early aughts? Office 2003 with XML and Sharepoint just wiped out all other stacks, especially Exchange steamrolling Domino and GroupWise.
I was part of a consulting team back then, the transformation projects were all going in one direction. That stack is dominant to this day.
I remember hearing that they had a storage calculator spreadsheet that would allow you to calculate how many IOPS you'd need from your storage to deploy Exchange.
Everyone I knew working with Notes or other products was gobsmacked. Why would you need that? Just buy decent storage. Nobody outside of the Exchange world cared about IOPS, unless you were dealing with of high thousands of users on one machine/cluster.
But the Exchange Server storage layer was a total dog's breakfast. (And not a healthy dog, either.) So disk performance was much more important for them.
And that spreadsheet was like crack for the senior management layer. They suddenly got to play with the numbers of users, mails per hour, mailbox quotas - then feel like they had involvement. They had input. They could go to their storage vendors with those figures and speak authoratitively about their requirements.
Those vendors would humour them, and sell them the exact same array that everyone else was buying for that capacity requirement and budget.
But Microsoft had turned a weakness into a strength, by making the managers feel like they had real input into the process.
Honestly, that spreadsheet was a genius move that it took me years to understand.
It's fine, email / photos app / texting apps are solved problems.
Keep upgrading the backends for compatibility / security / whatever, but you do not need to change the UX annually. We're done. Just stop.
Simply making your product worse is not normally a solution outside monopolies.. ah, I see.
I’m still searching for signs of 10x productivity in SWE showing up in actual consumer facing product improvements industry wide.
That and, you know, Microsoft.
Notes followed a similar pattern, with a proprietary "post office" system with optional gateways to larger x.400 / SMTP / UUCP / OfficeVision networks for those who needed them.
Going SMTP (or even x.400)-native from the start might have been a better choice in the long run, but that's only clear in retrospect - it certainly wasn't obvious at the time.
Even if your typical big company did have a gateway for their internal email to other companies, odds are very few people used that gateway for business. If you wanted to communicate with somebody in a different company you dialed the phone or sent a fax.
The Notes directory was based on X.500 - it had the same attributes etc. Notes didn't implement DAP, but then almost nobody did, which is why we got LDAP.
There was an add-on connector for X.400 mail. I remember installing it a couple of times. It was fine, if cumbersome due to the nature of X.400. ;-)
They also didn't ignore SMTP. There was a connector for it for Notes 4.0 and 4.1, and starting with Notes 4.5 that connector was bundled in the box. Starting with Release 5 they retired the connector and moved SMTP functionality to be native in the mail router process, and for security they provided a separate SMTP listener task. Both of these were configured in the Notes Directory via Server Documents and Configuration Documents, just like any other feature of the server. If a Domino server wasn't sending or receiving SMTP, it's because its administrator didn't want it to, not because it couldn't.
It wasn't a walled garden, or if it was the wall was very low. By comparison Microsoft's first attempt to dethrone Notes - Exchange Public Folders - was a completely walled garden. And a very well fertilised one which people avoided due to the smell.
But it's OK, Microsoft then gave us SharePoint. The fact that Microsoft's customers rushed to it tells you how bad Public Folders were. So your comment "That and, you know, Microsoft" is much closer to the truth than protocol support.
(Source: I was a Principle Certified Lotus Domino Administrator for multiple versions, and have used & administered Lotus Notes versions from version 3.33 through to 8.5. I worked at a consultancy and installed many, many implementations of Lotus Notes across a number of topologies. I also had an Exchange Server MCP and worked with multiple versions of that too.)
It's hard now to appreciate how different corporate IT was from the free-wheeling Internet world. The unexpected popularity of the web really did change everything.
Plus just how much SMTP sucked compared to other systems and only got enough kludges barnacled onto it in semi-transparent way by late 1990s