DEV Community

Brian Rinaldi
Brian Rinaldi

Posted on with TJ VanToll

4 Things Developers Take for Granted That Used to Be Really Hard

As software developers we love to complain. Spend a few hours in a development office and you’re bound to hear shouts of bad documentation, horrible error handling, and the how the previous developer was the worst.

And even though these are real problems, every once in a while it’s good to take a step back and consider how far we’ve come. Because, although the issues we deal with today seem hard, these annoyances feel quaint given the perspective of time.

This article is a nostalgia-laden guide through some advances in the last 20 years of software development that you’ve forgotten about or never lived through. Your authors for this journey are Brian Rinaldi, who is old enough to remember when IE6 was considered awesome (because it replaced Netscape 4.5.2), and TJ VanToll, whose career started helping others add background music to Geocities sites. You’re in good hands.

NOTE: You’re absolutely welcome (encouraged!) to contribute your own stories in the comments, especially if you have, ahem, more experience than your authors.

Thing #1: Actually Getting Software

Back in our day, if you wanted software, you had to get into a car, drive to a store and buy it off the shelf.

shrink-wrapped software
We used to get excited to buy this stuff
Source: cogdogblog [CC0], via Wikimedia Commons

If you are really old, like Brian, you remember the days when software came on sets of floppy disks. You had to install the software by loading each disk in the proper order.

floppy disk installs
Half of the quest was getting this thing to install
Source: http://retrothrifter.com/product/kings-quest-6-vi-adventure-retro-game-ibm-5-25-floppy-disk-big-box-1992/#&gid=1&pid=1

Or perhaps you are lucky enough to only recall when that software came on CDs or, heaven forbid, DVDs.

Software was expensive too! The full boxed version of Dreamweaver Ultradev in 2002 cost $599. Flash MX in 2004 was also $599.

The high cost of software and ease of sharing physical disks meant that anti-piracy measures were often incredibly burdensome.

Dreamweaver activiation
I don’t even know where to start here
Source: https://helpx.adobe.com/x-productkb/policy-pricing/macromedia-legacy-activation-error.html

As if activating by phone wasn't enough, software such as QuarkXPress actually required you to have a physical hardware dongle attached to your computer to function. Yes, young people, you read that right!

anti-piracy dongle
Bonus points if you know which port those top two dongles needed
Source: https://www.planetquark.com/2017/06/14/is-a-subscription-basically-a-dongle/#.W4BiJZNKh-U

Nowadays it’s hard to find software that you can’t find on the internet one way or another.

So, sure, today you are often stuck with perpetual software subscriptions, but we've still come a long way.

Thing #2: Web Site Hosting

The biggest problem with web hosting back in the day was that it was insanely expensive. Just basic hosting might cost you anywhere from $30-50 a month. Want a database? Well, a SQL Server or MySQL database may double the price. Want SSL? That’s another monthly fee, plus you’ll need to buy a certificate. Oh, you want to install that certificate? You’ll need a Ph.D. in cryptography to accomplish that.

Homer hard at work
Actual footage of Brian reading Apache’s documentation on installing an SSL cert in 2003

Sites like Geocities and Angelfire were wildly successful in their day because they provided free web hosting, which was unheard of at the time, and which allowed developers to share their <font> tags, <table> layouts, and amazing designs with the world.

Geocities
Imagine how good this felt in IE 5 with auto-playing Doom music.
Source: The Internet Archive: https://web.archive.org/web/*/http://www.geocities.com:80/Hollywood/2979/

Of course, there’s no such thing as a free lunch. And Geocities’ auto-injected ads, verbose URLs, and rampant popups (remember those) ensured that anyone developing professional sites would need to dump their money into more a more expensive solution.

pop up ads
That’s right kids, websites used to open whole windows that you had to close manually, like an animal
Source: https://www.engadget.com/2014/08/14/the-creator-of-the-pop-up-ad-says-sorry/

Nowadays basic free web hosting is a given. With GitHub Pages you can get a site off the ground in minutes, and services like Netlify even offer free SSL certs that you don’t have to install yourself. And speaking of GitHub, there’s one other massive change to the way we build software we have to discuss.

Thing #3: Version Control

If you wrote software in the late 90s or early 2000s, there’s a 100% chance you can tell a story about some horrible proprietary version control software you used. Or, perhaps, like many of us, you didn’t even use version control.

For example, if you’re like TJ, you recall using an external hard drive as “version control”, and prayed to your deity of choice that the hard drive would never be needed. Brian, on the other hand, recalls the “official” version control at several previous roles involved renaming files with “.old” or _old” in the filename. It might look like something along the lines of the example in this presentation by Gwen Lofman:

bad version control
Yes, we've all done it.

Needless to say, these methods were less than ideal. The thing is that legitimate version control, until fairly recently, might be free but not easy (things like CVS, SVN) or both not free AND not easy (Visual SourceSafe).

Learning to use these tools often took quite a lot of effort. Poor merge conflict handling might end up making your code look something like:

something is not quite right
Something isn't quite right here
Source: https://redfin.engineering/visualize-merge-history-with-git-log-graph-first-parent-and-no-merges-c6a9b5ff109c

Heck, even in the early days of Git, it had a steep learning curve.

The advent of GitHub (and other solutions like BitBucket and GitLab, for example) really changed everything. They took a lot of the initial pain of setting up a version control repository and made it a no-brainer.

Thing #4: Open Source

In today’s age, it’s hard to remember that just 10–20 years ago the majority of code developers used were closed source. This wasn’t limited to things like your your operating system, application server, database or desktop applications. Want a message board for your site? Take out your wallet. Want a simple CMS? Be prepared to pony up. Want a simple rich text editor for your form? That’s gonna cost you.

This impacted more than just the cost of the software. For instance, hit an issue? Maybe there was a place you could report the bug online. Maybe you just cross your fingers that the bug will be fixed in the next year’s version.

waiting on bug fixes
Source: https://devrant.com/rants/1314645/bring-the-funny

The other thing that is hard to imagine today is that many people were thoroughly skeptical of open source. For example…

In fact FOSS is often a recipe for half-baked freebies that displace good software because they are free, but which then go on to soak up everybody’s time by being full of bugs that require ‘support’ (i.e. your time or money) to heal.

Dr Mark Tarver, The Problems of Open Source

And companies, with Microsoft often at the forefront, fought open source intensely. Today’s Microsoft is one of the largest corporate contributors to open source and even owns GitHub. In fact, open source is so thoroughly ubiquitous in developers’ workflows that developers often tend to take a liberal open source license for granted (and, yes, this is a bad idea). Yes, I think that means open source has won.

Wrapping Up

Whether it’s open source, version control, web hosting, or the way we actually acquire software, a lot has changed for the better in the last 20 years. In this article we looked at a few of the changes your authors have appreciated, but of course we’re only scratching the surface of advances in the technology world.

If you’d like to contribute your own item to this list let us know in the comments 🙂

Top comments (67)

Collapse
 
raymondcamden profile image
Raymond Camden

Instead of a software key, many old games required you to own the docs and would reference them in game as a way to ensure you hadn't stolen the game. Ie, go to page X line Y and type in the word.

Yep - don't miss that at all.

Collapse
 
blazselih profile image
Blaž Šelih

Oh yes. This brings back memories.
And sometimes docs were printed with light green ink on dark green background to make them impossible to photocopy.
This was, of course, before the wide spread of scanners and cheap color copiers.

Collapse
 
raymondcamden profile image
Raymond Camden

Although to be fair, some of the stuff you got in games in the old days were incredible, especially Infocom.

Collapse
 
remotesynth profile image
Brian Rinaldi

Omg. I forgot about that! Yes.

Collapse
 
hagailuger profile image
Hagai Luger

You've just taken me 25 years back!
Paging through the documentation to find the right word (and probably copied the documentation too...)

Collapse
 
aeiche profile image
Aaron Eiche

I remember this in Prince of Persia on a Macintosh LCII in middle School. Not having the instructions, we'd guess. Occasionally we'd get it right.

Collapse
 
bgadrian profile image
Adrian B.G.

First I needed to rollback the tape when a game loading stage failed.

I still remember helping my father installing Windows 95 from 13 floppy disks on 10 PCs at a school laboratory, and I had fun!

Now I do docker run and I can have any OS or software in a few minutes. Amazing.

Thing #5: learning resources

From books to courses, from slack to stackoverflow, we now got it easy and free.

Thing #6: Browser dev tools and HTML5 standards in all browsers

Specific to web devs, but nowdays is very easy to start with front end development. Although it is the most difficult job to master I would say.

Collapse
 
elmuerte profile image
Michiel Hendriks

Oh yes, thing #5. One of the reason for those "big boxes" of software was for the huge amount of dead tree content called: manuals. Back when the fancy new thing called "online help" was a digital manual you easily open from within the application jumping directly to the relevant section.

Collapse
 
bgadrian profile image
Adrian B.G.

Oh yes, the help manuals files ... Good ol times

Collapse
 
remotesynth profile image
Brian Rinaldi

Agree on both!

Collapse
 
merrythemes profile image
Merry Themes

Rollback the tape

That brings make some bad memories.

Collapse
 
imthedeveloper profile image
ImTheDeveloper • Edited

I literally laughed out loud at some of this..the memories. I had windows 95 on about 50+ diskettes it wouldn't install even on a 486 after getting up to number 31. I could have cried that day if I wasn't so battle hardened.

I spent a lot of time on angelfire, Geocities, homestead and begging Dev friends if I could take some space on their web servers. Countless times my site would go offline due to providers going bust and I remember spending plenty of time adding scripts to remove forced adverts and branding all the free hosting providers put on your pages

Collapse
 
ntnunk profile image
Noel

I remember getting the original "Master of Orion." It came on, if I recall correctly, 10 1.44" floppies. Me and my buddies at the time where aghast and all we could think was "What in the world would require any game to need THAT MUCH DATA?!?!?" Times have changed for sure. These days I have a glorified text editor (Eclipse) sitting on my hard drive and sucking up 68GB...

Collapse
 
guneyozsan profile image
Guney Ozsan

I remembered playing Monkey Island 2 on Amiga (no hard drive and like 10 floppies). In one scene you needed to switch disks when a character spits, and you needed to switch back the disks when the flying spit lands on your face. It was so hard to keep it going without a HD.

Collapse
 
david_j_eddy profile image
David J Eddy

Geocities was the reason I found out about CSS :D. z-index FTW!

Collapse
 
simonhaisz profile image
simonhaisz

Code reviews. Even when using version control 10+ years ago the code review would be you sitting at another devs desk as they walk you through the diffs.

Continuous Integration. Remember having to commit your changes in the morning so the noon build would get them? And because the build takes 2 hours and the first set of regression tests takes another hour you only have a few hours to fix any issues that happen before the 6pm build?
Yeah, I don't miss that.

Collapse
 
shawnoden profile image
Shawn Oden

Sadly, I recall too many times when "continuous integration" was just Ctrl-S, since there wasn't anything other than the live site. And QA was just reloading the website.

Collapse
 
rubberduck profile image
Christopher McClellan

Why aren’t your code reviews like that now?! Absolutely the most effective way to perform a code review, in my opinion. In fact, it’s so effective that my team does it continuously...

Collapse
 
simonhaisz profile image
simonhaisz

I believe that there are pros and cons for both types of code reviews, which I will refer to as sync and async. Depending on your situation one may work better than the other for you.

sync
Pros:

  1. Everyone is focused 100% on the review
  2. Questions can be answered immediately - "I don't know" is a valid answer
  3. No context switching during the review means efficiency That's all great, so even after we got software that supported async we kept doing sync. But then we ran into the following cons so we switched.

Cons:

  1. Requires reviewers to find time when both of them are available at the same time.
  2. It can take a while for the reviewer to process the changes so that they can even start asking questions. All the while the writer is just sitting there waiting...
  3. ...Which lead to their frustration sub-consciuosly influencing the reviewer to 'hurry up and approve' without them saying a word. During this time we had a number of bugs crop up that were the exact types of issues that code reviews are best at catching. Our RCA of this was that the nature of sync reviews put more pressure on the reviewers to approve changes quickly, meaning they were disincentivised to go "I'm not sure, I need to think about this for a bit". Now, depending on your team/culture maybe this doesn't happen to you.

What we've experienced from async
Pros:

  1. Allows people to review on their own time/pace, including remotely
  2. Not having someone waiting next to you while you review means you do not feel rushed (any more than the current schedule anyways)
  3. Documented feedback means a) things are likely to be followed up b) others can learn from the review
  4. If the reviewer needs to have a face to face discussion, they can still do that! It's not uncommon for me to add a comment "I'm confused/concerned about this approach" and then meet with them
  5. It's easier to review code not in the review using my own machine/setup rather than all to drive on their box.

Cons:

  1. Promptness of response can really lag
  2. Someone can technically click the green check mark without looking at a single line of code Depending upon the maturity of your team those may be big issues or not issues at all.
Thread Thread
 
rubberduck profile image
Christopher McClellan • Edited

I was just making a joke about pair programming man. I’ve been on teams that don’t do code reviews anymore because we pair anyway.

Thread Thread
 
simonhaisz profile image
simonhaisz

Ugh woosh right over my head...

Thread Thread
 
jrohatiner profile image
Judith

Very thorough explanation. And accurate as well. Namaste!

Collapse
 
mazz profile image
Michael • Edited

I don’t recall ever working at a place with that high a degree of build integration. Maybe one place but that was mid-to-late 2000s

Collapse
 
simonhaisz profile image
simonhaisz

That example WAS from the mid 2000s :)

Collapse
 
remotesynth profile image
Brian Rinaldi

Yes. Agree on both.

Collapse
 
theycallmenucci profile image
Steve Antonucci

Somehow I feel like intellisense (and maybe even Google? 😁) belongs on this list. Prior to that we had to buy, store, dig and sift through very large reference books in search of functions and their signatures. 😂😂

Collapse
 
bokmann profile image
David Bock

I still have a shelf in my basement filled with the ‘Inside Macintosh’ volumes. There are literally like 20 books. The legendary one was Vol. 6... I’ve seen smaller sets of encyclopedias!

Collapse
 
theycallmenucci profile image
Steve Antonucci • Edited

LOL! I have a similar problem, still have lots of old books from the 80s and 90s. I have to just force myself to accept that they're just never going to be interesting to my kids and they'll never have any meaningful historical value, despite how hard I want either of those to be true. 😂😂

Collapse
 
ferricoxide profile image
Thomas H Jones II

If you are really old, like Brian, you remember the days when software came on sets of floppy disks. You had to install the software by loading each disk in the proper order.

If you're really old (or started playing with computers as a kid), you remember some software (games, mostly) coming printed in magazines and having to transcribe it to your pre-PC (Trash-80, Apple ][, Commodore, etc) and then save it out to casette if you didn't want to have to type it all back in again. And you still clearly remember having to cue the tape up to just before the screaching so that when you ran the load command, it would actually be able to read data before the load-command timed out. :p

Or, perhaps, like many of us, you didn’t even use version control.

I still remember the pure joy, in 1989, when a college friend pointed me at the most wonderful tool I'd encountered up to that point, rcs. And, in the non-UNIX wastelands of 8.3, you weren't couldn't even engage in:

file1
file1.bak
file1.bak.2
...

In today’s age, it’s hard to remember that just 10–20 years ago the majority of code developers used were closed source.

Though, before the advent of SysV- and then Windows-based systems, things were amazingly open. I still remember the shock of when they upgraded our Sun lab from SunOS 4 to Solaris and, suddenly, cc was something you needed to request access to (because they suddenly needed to buy compilers).

P.s. I'm old, apparently (and starting my computer journey at age 7 has made that journey even longer than just the span of my career).

Collapse
 
simonhaisz profile image
simonhaisz

Setting up environments.

Today we have VMs which are quick and easy to spin up and then put software on. Or even better, containers that already have the software you need and cannot be changed!

I remember having to use ghost software to re-image machines every morning to make sure that my environment was clean before installing last night's build. And those machines would be physically in my cubical, with a KVM switch to control them! Spinning up a VM and rdp/ssh into it? Way better.

Collapse
 
remotesynth profile image
Brian Rinaldi • Edited

OMG yes! I remember the days when not only was every dev's local environment was different, but dev was different from QA was different from prod. Made for some eventful deploys!

Collapse
 
simonhaisz profile image
simonhaisz

Exactly!

"Works on my machine" hasn't been fully killed off, particularly for client apps. But for a lot of cases it is as dead as the dodo.

Collapse
 
ntnunk profile image
Noel

It's depressing how much of that I remember first-hand (roughly all of it). And OMG, hardware dongles. I completely forgot, or maybe suppressed the memory of, hardware dongles.

Great article and a serious blast of nostalgia along with my coffee this morning. Thanks for sharing!

Collapse
 
desolosubhumus profile image
Desolo Sub Humus 🌎🌍

I remember pulling out a book and having to use it to help with writing the Fortran software to make the Hangman game, then save it to cassette. The old Classic99 games* were a huge upgrade from that, and floppy disks were something I wouldn't have until sometime later on.

The worst? Wanting to play music with the cassette player and hearing the software you wrote play.

*mainbyte.com/ti99/howto/classic99....

I feel really old, now.

Collapse
 
desolosubhumus profile image
Desolo Sub Humus 🌎🌍

Bad me, replying to myself ...

I just saw Google Tone in the Chrome Web Store. It's like the old cliche 'everything old is new again'.

I mentioned how saving software on a cassette tape ended up saving data that a tape player translated into sound above. Theoretically, with the right equipment and software, you'd be able to translate the sound back into the original code. I'm guessing Google Tone works on that principle, but plays URLs instead of Fortran, like my old cassette tapes did.

Sorry all, I couldn't help myself and had to share that.

Collapse
 
govcoder profile image
Eric Wilson

Debugging! Cross browser compliance is a breeze compared to figuring out what was causing a Windows 95 desktop application to break on that 1 user's machine. What was that dll every software company on the planet wanted their own version of ctl3d.dll or something?

Also, stability 20 years ago everything randomly went down all the time and seemed to have very little redundancy. Desktops crashed often (blue screens), network connections dropped frequently, if someone knew your IP Address it was trivial to crash your workstation (ping of death, etc...).

Collapse
 
theoutlander profile image
Nick Karnik

I installed Win 3.0 probably over a hundred times! I think it was ~13 diskettes. I did that for a ton of family and friends too! Most of my life was spent fixing corrupt machines and reinstalling everything.

Followed by Windows 3.1, Windows 3.11, Windows for Workgroups, Windows 95 (which was 21 disks I think).

Collapse
 
spidergears profile image
Deepak Singh

Thanks Brian for giving us perspective from back then. Now we have come a far way from back then, documentations are just a google search away, OSS just a github search, an awesome shared knowledge-base is at stack-overflow, to name a few. This is awesome time.

At the same time I see a lot of people have become complacent with abundance of these resources, which a cause of concern.

I was a student when CDs and Floppies were still in circulation, we would get programmes and games on CDs and use floppies for data needs. I was confused when I saw a thumb-drive for first time, how can it be so small and still have a huge data capacity compared the 3 1/2 " floppy.

Collapse
 
moopet profile image
Ben Sinclair

We got bored of printer-port dongles at university and user a manual printer switch.
I remember cracking a game I bought because it was missing the manual and the copy protection was of the kind where you had to enter word such-and-such from page n of the manual.

These efforts were easier than going the "legit" route. Copy protection back in those days was arduous.

Collapse
 
superdiana profile image
Super Diana

OMG this is amazing! lol. I did use both 5 1/4 and 3 1/2 disks! - i think the first game i installed on my own was california games or maybe prince of persia... lol i was precocious!. Totally loved being part of the generation that saw new things like internet and mobile phones. I can definitely say back then nobody imagined we would get this far.

Collapse
 
ozzyaaron profile image
Aaron Todd

The first things that often come to mind is

  1. Having a screen & keyboard to work with.
  2. Being able to look at your screen for a while without getting a headache.

I started shortly after screen & KB became the norm so we still had a class or two demonstrating older technologies. I feel that it is easier to translate, explain and get your ideas working with currently input/output.

On screen technology when I started I think we had 12" CRTs and I definitely remember 14" LCDs with ghosting and so on at the beginning. Even with ghosting though you almost instantly were able to code for 8 hours straight without eye strain. There was no way to do this with a CRT - even the very best and most expensive Sony ones :)

Some comments may only be visible to logged-in visitors. Sign in to view all comments.