Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Thunderbird Time Machine: Windows XP and Thunderbird 1.0 (thunderbird.net)
94 points by HieronymusBosch on Aug 18, 2022 | hide | past | favorite | 65 comments



I don't know if it's nostalgia, if it's XP being the pinacle of user interface design, or if it was the thunderbird designers, but the screenshot looks better than current thunderbird. The balance between the items and how much that ask for your attention feels just right


It's not just nostalgia:

- the buttons are big

- there is high contrasts

- sections are clearly delimited

I have the current Thunderbird open all day, and I'm constantly searching for things. It looks much better, much leaner and cleaner. But nothing stands out from the rest, everything is a shade of grey, buttons are flat, section are loosely separated or sometimes merged, labels don't have icons.

It's like the debate about cars with touch screen or buttons: the touch screen makes for a much better look, but its usability sucks.

It's not just thunderbird. Phone apps, web apps, they all follow trends that have nothing to do with usability. They copy each others in loop, bored designers go one more round (see ubuntu for an extreme example of this), A/B testing tells manager people downloaded your app more because of the pretty UI or screenshots on the short term.

But the current design is a usability disaster, and is only partially working because the UI are getting dumbed down more and more. Apps have less and less accessible buttons and menus. It makes you less productive, but it's easier for the average user, thanks to the paradox of choice. Because of this, having a lean UI is not as terrible, since you don't display much on the viewport anyway (e.g: youtube latest changes, or just tik tok). But of course, advanced users pay a price for that. Also, you can't do it for professional software unless you want to render it harder to use. Case in point: Thunderbird.


We used to have artistic masterpieces for icons compared to the near incomprehensible "silhouette" style uninspired hieroglyphs we call icons today.


To me the 1.0 screenshot looks nearly identical to the Thunderbird window I have open right now.

I don't have email-related buttons, since I have no email account configured. I use Thunderbird for RSS feeds. But the rest of the interface is exactly the same, except that the window topbar is tabbed in modern Thunderbird.

As for Windows XP being the pinnacle of design, I always hated its default theme (visible in the screenshot of the setup window, though, oddly, not in the screenshot of the running app) and changed it to the Windows 95 theme on all of my computers. I still hate the XP look.

EDIT:

I was wrong; I do have the email-related buttons. I've just never noticed them.

Instead of square icons, they are now text labels (with pointless icons next to the text in most cases). I don't really see that as a problem. Usually the problem in modern design is that the pointless icon is viewed as a substitute for the text.


Windows 2000 was the pinnacle of GUIs. Things have gotten worse since then. In the past five years they've fallen off a cliff. Absurd amounts of whitespace. Buttons without borders. Icons only instead of text buttons. And I have to reiterate the absurd whitespace quantities.


I would call the design aesthetic du jour "the White Wasteland". It's a symptom of what I'd call "fundaminimalism". Minimalism isn't automatically better when it comes to user interface design, but too many designers seem to think it is.

On a related note, why does everyone have to slavishly imitate Google's ugly, bland, undifferentiated interfaces, why?!? Nowadays, GUI seems to stand for Google-Ugly Interface.


> the White Wasteland

> fundaminimalism

> Google-Ugly Interface

The irony is that you put more effort into coining those pejoratives than the UX designers put into their UIs.


I’ll be honest, I liked the XP "Fisher Price" look even more.

Yes it was hype to mock its colorful style, but it’s forgetting that every UI control was refined, consistent, with enough relief and contrast but still sober enough to not steal the app identity.

It was flashy for sure but for me it was a breadth of oxygen after decades of grey. Icons were very discoverable, Control Panel was just the OG control panel but sorted into categories.

It was a beautiful OS for the present, not an UI coming from a futuristic movie like everything today where animations and great looking empty space are more important than usability.


Exactly this. Aero was great not because of itself, but because it was basically Fisher-Price Windows 2000, where things were clearly things.

The Mac from around that time was similar, even if pinstriped.

Touch and the anti-skeumorphism has destroyed computer UIs.


IIRC, Aero was the Windows Vista UI disaster, while Luna was the reasonably usable (albeit excessive) Windows XP theme.


if you like windows 2000 as a pinnacle of GUIs, you should be quite happy with the current stable versions of XFCE4


I've been using xfce for almost a decade.


Agree with you there.


> pinnacle

It was nearby. The class of NeXTSTEP, Win2k, SGI, and aspects of the Amiga/MacOS9 around the turn of the century are some of the best computing interfaces ever made.

Intuitive, direct, consistent, easy to learn & use. Labels concise. Apps respected system themes with the exception of a few "skinz" media players, and they were even customizable! Gentlemen, we have the technology.

I'd prefer one of these with the addition of modern niceties like menu search. Mate is close, but it is slowly rotting via GTK3/4 from the bottom up. I.e. the Disks app on a 4k monitor is a mess.


I think KDE still has the style of older systems, or at least to me it seems like it does. GTK DEs are all a total disaster and definitely a sinking ship for anyone wishing to have any customization or usability over their system, I booted up XFCE the other day on Linux Mint, and 3 different programs all had their own title bars. Things like that don't ever seem to happen on Windows or KDE.


Yes. KDE however suffered for many years from what I like to call "programmer art." Including bad fonts and padding. It also had a bewildering amount of control panels and tweak-ability.

Better than a lack of them, but Win2k had it just about right. A reasonable amount of configurability, mostly good defaults, and just about anything possible via a setting hidden underneath.

Gnome and KDE seem to have chosen opposite sides of that happy medium.

Looking at recent screen shots KDE seems to have fixed the first problem. I'll have to try it again.



Neat, but these don’t have the directness that even Win 3.x had, click on a widget and set the color.


I'm still mad at whoever decided to hide the entire menu bar in a burger menu and made everything in it take twice as long to find and use.


When they take new designs to user research, one of the questions asked is "Which one of these options looks 'cleaner' in your opinion, A or B?" And when one of the objectives of the org has been declared as "clean up the UI" I'm sure you can imagine why we end up with the disgusting hamburger menus everywhere. And the people who championed a 'win' from such a change will parachute to the next company and do the same thing.


If it's like Firefox, the menu bar shows up when you press Alt.


And with two clicks you can make it permanently visible.


Took me two years to accidentally find this. That's a usability sample.


I'm think it's a combination of nostalgia and familiarity. Those were the kinds of UI's I used when I took classes about computer programs and spent a lot of time independently learning programs.

All those hours spent makes it feel more comfortable though it may not actually be better.


I think ""preferred familiarity" over time is literally the only meaningful definition of "better."

I'm growing quite tired of the deeply stupid idea that these UX/UI people understand how I work better than I do.


"Preferred familiarity" is a subjective thing, and since you've developed your preferences there have been a billion new computer users who are more familiar with the modern designs. That's why the designs change, things like Thunderbird aren't the trend setters.

Ideally both UI gets maintained, but that's not always feasible.


Sure, but lets keep clear what you're saying:

"Despite the existence of working preferences, so called modern-designers spend their days changing things for potentially no reason -- but even if there is a good reason -- they're also breaking backward compatibility."

In other words, let's not pretend how design-fashion is executed is always good or even neutral in operation in terms of helping people.


>let's not pretend how design-fashion is executed is always good or even neutral in operation in terms of helping people.

This isn't unique to modern design. Your preference is also full of bad decisions that don't always help people. You've spent decades using them and are now accustomed to the way they work and don't even see them as flaws.

It's largely familiarity, and is another form of "things were better back in my days."


Come on. Please don't pretend that there aren't severe institutional defects now in how design-fashion is executed that go beyond "sometimes old is better, sometimes new is better."

Large centralized purveyors of software with monopolistic-type power and no liability are generally (perhaps subconsciously) are frequently incentivized to push design into negative directions, e.g. dark patterns.


I understand why you think these things, but at the end of the day many of these new designs are quite popular.

For example, every time I even momentarily end up on the new Reddit, I think "how could the designers have approved this cosmic mess?" Yet presumably people like it, or at least it's profitable as I doubt they'd stubbornly hold onto something for four years if it were driving users away.

Remember the old "utility first" brought us things like Clippy too. And the generation before mine also sees all that GUI nonsense as cruft too. Richard Stallman talks about how he still does basically everything in emacs.


You're implying that they're popular because they were chosen by users.

Often times, they weren't. They were effectively forced on people.


I'm implying their popular because they're familiar, that was how my first post started. People get comfortable with things forced upon them too.


> since you've developed your preferences there have been a billion new computer users who are more familiar with the modern designs. That's why the designs change

That is logically impossible; the designs cannot change due to the emergence of users who are more familiar with future designs than they are with present designs. No such users can ever exist.


It's only impossible if there's no other software.

The users at Thunderbird's release were familiar with Netscape and Outlook so the design emulated them. Any new user is familiar with Chrome and Gmail, so Thunderbird moves towards designs the new users would be familiar with.


Then you're arguing that Thunderbird is trying to emulate Chrome. But you haven't even attempted to argue why the design changed. Chrome was not facing a population that was familiar with designs from the future.

> The users at Thunderbird's release were familiar with Netscape and Outlook so the design emulated them.

If this were true, then Chrome's designs wouldn't have changed. You can't have this both ways.


Outside possibly contradictory influences are a thing. Like much of modern software is influenced by the rise of mobile, where the different limitations necessitated different designs.


I prefer Win8 (or non-initial Win10, initially it lacks color on title bar IIRC), it's well colored but not disrupting by transparent.


What's interesting to me is that I actually like the old interface better than the current one. But the part that matters, the actual email list, has barely changed in 20 years. This is the place where I feel they need to modernize most. Thunderbird needs a robust, clean conversation view, like GMail or macOS Mail.app (That app is actually one of the main reasons I use a Mac right now. If there was an on par OSS version I would probably use Linux.)

Also, I miss proper Outlook (Office365/Exchange) integration. I hate it, but if you use it at work it has to work. There is an integration, but it is a paid plugin.

That said, I really hope Thunderbird succeeds and becomes relevant again. If you have a certain amount of mail, a real mail client is just so much better than webmail, and it is great to have an OSS version.


> conversation view

One webmail I use has this, but shows only one line per message, with giant headers. I have to expand the thread anyway so don't see the point. Maybe if it could show most of each message like a chat client it would be useful.

> becomes relevant again

Been using it since it was called "Netscape Mail."


> the actual email list, has barely changed in 20 years

That's about to change:

https://developer.thunderbird.net/planning/roadmap


I'd pay for a modern (secure) Eudora port. It was a great software and 7.1 source code is available https://computerhistory.org/blog/the-eudora-email-client-sou...

There was a Thunderbird Eudora client (well it was an extension) but got abandoned quickly https://en.wikipedia.org/wiki/Eudora_OSE


I still use Eudora as my primary client for both home & work email use. IMAPS, POP3S, TLSv1.2, etc.

The biggest help in accomplishing this was a project from few years back... https://sourceforge.net/projects/hermesmail

You basically need 3 DLL & 1 P7B files to patch the default 7.1.0.9 install to gain the required functionality for using a modern secure mail server.


Not related to Thunderbird but I think it's a testament to Microsoft's backwards compatibility that you can just run a program from 2003 with no issues whatsoever.


See, I'm torn on this. Part of me feels that the ability to use older hardware and hence older software (without being forced to upgrade both HS and SW) is liberating, fun, and gosh darn useful! ...And it helps with folks who financially lack the means to constantly keep upgrading. But, then, the other side of my brain says: there are some legitimate reasons for *some* upgrades like security updates, simpler ways of doing the same processes, etc. Also, developing software without needing to be too backwards-compatible does make it so many more people can participate in development without too much burden. I imagine some yound person new dev and full of creative new ideas need not to learn tons of legacy code only for backwards compatibility. That approach of freshness can be liberating too!


Let's not exagerate. You can run some programs, not all.


Still beats Gmail's UI/UX 20 years later


What’s email provider got to do with a software provider?


> Visually, it prided itself on having a pretty consistent look across Windows, Mac OS X, and Linux distributions like CentOS 3.3 or Red Hat.

That’s probably one reason why it never was very successful on OS X.


yes, I recall eagerly installing Mozilla + Thunderbird 1.0 in 2004 but on Hoary Hedgehog (Ubuntu 5.04) not WinXP.


I'm all for nostalgia, but I can't say I recommend putting your email credentials into 20-year-old software running on a 20-year-old OS connected to the internet.

EDIT: why the downvotes? Windows XP can be infected with malware merely by plugging in a network cable with no user action. If you follow this blog post you're likely to lose your account. https://security.stackexchange.com/questions/185642/how-can-...


Target audience is likely running it in a VM and behind a NAT, if they even have a copy of XP. SP3 wasn't as bad as folks remember, look into "slipstream" for the install.

Also, there are not swarms of infected XP machines around to attack as there were back in the day. Likely no active attacks for TB 1.0 as well.

Worst case something might sniff an email password from an old version of SSL. Don't log in to important accounts I'd say.


>why the downvotes?

The blog post wasn't recommending putting your credentials into ancient software. It said if you're feeling adventurous you could fiddle around with it yourself by downloading it from the archives.


Perhaps, but I can't imagine how you would fiddle around with an email client without adding an account, unless you just wanted to look at an empty list.


The few interested are already willing to go through the trouble of setting up a VM with an ancient OS. Presumably they could find some old database or virtualize a mail server for it or some other workaround I'm not imagining.

I doubt anyone is going to read that and say "what a good idea, let's install Windows 98 on my hardware and muck about."


If memory serves correctly, in the Thunderbird 1.0 era, all of the passwords would have been saved via encryption via NSS's secret decoder ring methods. So if you set up a master password, then the contents would be encrypted on-disk. (Even if you didn't, they're still not stored in plaintext--I just think it's encrypted with something like an empty password instead, but I'm not certain.)

Additionally, I'm pretty sure even 20 years ago, Thunderbird would have still thrown up resistance at trying to connect to an email server without using STARTTLS or SSL.


I wouldn't be worried about encryption. I'd be worried about any of the hundreds of widely-known vulnerabilities that were patched between 2002 and today.

https://www.cvedetails.com/product/3678/Mozilla-Thunderbird....

https://www.cvedetails.com/product/739/Microsoft-Windows-Xp....


Yes but there is a difference between possibility and probability: it is technically possible to hit these vulnerabilities, but the chances are practically zero, especially in the context here where someone might try it once like mentioned in the article. It isn't like there are many malware authors out there trying to infect people running 20 year old software (outside of targeted attacks).

I think you are more likely to be affected by a vulnerability bug in some random modern macOS application with its own autoupdater than a vulnerability in a 20 year old Windows program.

Remember that these issues existed when that software was actually mainstream and yet the overwhelming majority of people wasn't affected by them during the peak of their popularity.


The difference is back then, the vulnerabilities weren't even discovered yet. Today, metasploit scanners are running 24/7 scanning the entire IPv4 address space for thousands of vulnerabilities at a time - including automated chaining of exploits through e.g. routers with misconfigured UPnP. You don't need to be targeted. As mentioned above, you can be exploited with no user action merely by plugging in an ethernet cable.

But then again, I'm not your mom. If you really want to do this I can't stop you. Go ahead.


I wouldn't worry too much about encryption at rest, if nothing else then because it's only being stored on a local device that you control. What I would worry about is that those credentials are then going to be sent to a server and a system of that age probably doesn't speak any modern version of TLS.


As opposed to say, logging into Google?


If you're a ProtonMail user, you could run the Proton Bridge on a Linux system in your network, then use PuTTY on Win XP to forward the ports from there to Windows. Then Thunderbird is only sending bridge-specific passwords unencrypted to localhost, and it's all modern encryption from there on out.


Oh what's the worse that could happen! :)


Your private data gets accessed by a random person instead of by a megacorp.


Same thing happens whenever you use any of those alternative "frontends" for sites like reddit, youtube, twitter etc.


Twas sarcasm




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: