> “So, how was it to use a WYSIWYG web page editor from over 20 years ago? Quite pleasant, actually.”
The dirty secret of web apps is that we’ve mostly gone backwards in usability compared to native desktop apps from 25 years ago.
Web apps are a mishmash of paradigms. Pieces of desktop UI are reproduced using a woefully limited framework inside a static request-based page navigation model. The user never quite knows whether an action will trigger a multi-second page refresh, whether the back button does anything useful, etc.
Desktop UIs had professionally designed human interface guidelines based on decades of actual research. On the web, designers are primarily graphics artists who pick fonts and pride on making buttons look like nobody else’s buttons. Icons are nowadays tiny monochrome line scribbles without labels. Just pray there are tooltips so you can figure out what happens if you press one of these icon buttons in a web app. (Or maybe it’s just an icon and not a button? No way of knowing, since the conventions that made buttons obvious have been thrown away.)
The web is the worst application delivery platform of the past 30 years, so of course it’s the one we got stuck with. Worse often wins by its simplicity and ubiquity. Everybody could author a HTML page and some gradually built their skills towards apps. This review of old Netscape Composer reminds of how important that was.
Ugh. Not this HN trope again. "Web is bad, the old days were good, designers are the worst, developers are the best"
Lazy, inaccurate, and offensive.
How is this limited to the web? 'The Desktop' suffers significantly more from not having a single, understandable paradigm. Is this UI modal? If I click this will it 'pop' me to another view? Or will it open a new window? Is that window represented in task bar/switcher/dock/whatever? How do I get back to where I was before?
Not to mention the issue that both macOS and Windows suffers from being in the middle of a UI-kit transition, with a bunch of half-and-half apps across the system, with the new system being of... dubious quality (I'm staring at you, new macOS Settings)
I think overall, I think there is a positive trend for usability of software on both the web and 'desktop' (whatever that means). But still, just like 20 years ago, people can still make bad software.
Edit: lol - just now as I'm trying to uninstall some AMD drivers, I get a dialogue box asking me to visit a long Microsoft Store URL, and it's neither clickable nor copyable. Desktop is perfect!
The demands on UX and UI have fast outpaced the ability of designers and their software to deliver on it. As new requirements come along, we are forced to make different trade-offs to those we made in the past, and perhaps that leaves those of us who don't need the new requirements feeling like the software has "got worse", ignoring the benefits to others.
To take one example, 99% of monitors in 1997 were 1024x768 CRT with single pixel density, which freed designers to assume that's what their end user was on. No such standard exists today.
Even if parent was a bit stuck in a rant (which I'm guilty of) I think a few forces were at play during the 90s desktop apps. Limitations made apps more focused on bringing rapid enough functionality in a stable manner. Web does offer too much rope to hang your users with.
My bet it that, if computing doesn't evaporate in the next decade, there will be a middle-ground solution.
While we have more option when building app for desktop, the reason the web took over is because of the fragmentation of desktop experiences (+ potential complexity required to install desktop app) so overall, it seems it was more important to achieve interoperability and simplicity of deployment than powerful UX. Even in recent history, if you take Figma vs Sketch (which was a serious contender to Figma in terms of features and UX) Figma won because interops + no need to install.
I agree, and I do have an addition. As a software engineer it some times kills me that product development is all done through Figma.
* Professional opinions drawn in higher ranked stakeholders whims
* It drown visibility and attention to other processes of software development such as infrastructure, integrations, etc.
However, that also gives some nice things: Stakeholders and designers can fight in their own corner.
Desktop business apps maybe looked better - at least those few that were not designed by the same engineers writing the code - but they all were a lot harder to deploy and maintain. You'd be happy that someone finally bought your app, only to discover that each department in the company has a different version of OS/hardware, with different network & security rules, and people installing all kinds of private stuff on their machines and messing w/ settings. A nightmare. It would take days for us to deploy new version, only to find out that half of the printers stopped printing for some unknown reason. Great, let's waste another week debugging that, while clients are screaming at us for not being able to do their work.
When I fully switched into the web dev world around Y2K, it felt so liberating. Never regretted it for the second...
Of course you'll get shit when you have people who have no background or experience working on user interfaces. We came few cycles into that already, ever since web democratized access to doing such things. There was even Mystery meat navigation as a term way back in 1998: https://en.wikipedia.org/wiki/Mystery_meat_navigation
This is nothing new. Amateurs going pro. DTP saw something similar, video as well.. lots of things.
What's new however is in mobile age, what are you even designing for? It's a moving target between device aspect ratio and screen sizes.
You're mostly right, but there were absolutely terrible native apps 25 years ago. Consider a WinForms form-filling app, with basically no organization of form fields, except that each one of them is surrounded by a different color of angry fruit salad.
> The dirty secret of web apps is that we’ve mostly gone backwards in usability compared to native desktop apps from 25 years ago.
I wasn't aware it was a secret! Though it certainly isn't just web apps: the desktop has taken steps backwards too, and not just where bad web apps are hosted in simple desktop wrappers.
Back in the 90s and 00s I remember a strong push for consistency and discoverability, where now that often takes a back seat to aesthetic matters and shiny new things. And it isn't just the apps: whole OS UIs have fallen apart, with levels of inconsistency well beyond that which Mac and Windows people used to poke fun at the Unix world for. It is amazing how hard it can be sometimes to see at a glance which app currently has input focus on a Windows desktop, particularly with multiple screens.
There are several parts to the consistency problem, and the overall UX-is-more-shite-these-days problem, including:
1. The conflicting priorities of the range of devices (trying to make something work on a small touch-based mobile and a large desktop environment, often creating something that is imperfect on both).
2. The desire to be subtle and pretty, even where this makes things less obvious (some flat designs are particularly bad for this: little indication of what is actionable, scrollbars becoming single lines of pixels or invisible entirely, title bar text being slightly-whiter-than-mid-grey when active and slightly-darker-than-mid-grey otherwise and sometimes the reverse).
3. The desire to have an app that is easier to monetise where a plain web-app would be better (so you get something that is trying to optimise for four distinct UX arenas: web, mobile web, mobile app, desktop app).
4. The desire for everything to be online-first, so you can more easily justify charging for a subscription (and so you don't have to worry about sync issues inherent in offline-first design) even in cases where entirely offline would be the better choice for the user.
5. The desire to be hip and new with UI design completely overriding years of earlier UX research because form sells better than function. Yes, some of those research results may have been questionable, and some that weren't are less relevant to today's environments, but they all seem to have been thrown away in many places.
It isn't all bad though, some things are still well-designed from a UX standpoint, or if not well-designed at least not brain-dead. I live in hope that people will learn from the few good examples, and we'll have a swing back towards caring about properly optimising for UX instead of just looks & monetisation.
General usability degradation is absolutely the experience of those who grew up before the smartphone revolution. Those who came later, one must presume, don’t know what they’re missing.
I'm not denying that perception, but as with everything that has ever changed, there is a factor of resistance to change and idealization of the past to it.
The audience of apps since smartphones came up (let's say 20 years ago) has vastly diversified, from previously mostly professionals and enthusiasts, to a far more general audience. That adds new requirements, and shifts priorities - it's more important now that simple things are simple, and it's less important that difficult things are possible at all. And from an inclusion pov, that's a good thing!
I'm not saying that everything has strictly improved, nothing ever does. But it certainly hasn't gotten strictly worse, nothing ever does either. It's just the tradeoffs that are different, and no longer target as specifically those folks who tend to yell at clouds on HN these days.
The issue is that it has degraded even in software that is purely targeted towards professionals and enthusiasts. And it’s not just the UX designers’ fault, the UI tech stacks have degraded in that respect and make it difficult.
I’m also hard-pressed to come up with anything that has improved on the desktop.
Oh, how I miss the Windows 98 UI. It just now occurred to me how genius the use bevels was. When glancing at a window, it is immediately clear what cannot be interacted with (flat things), what can be clicked on (bevels outwards) and what is an input element (bevels inwards). It's like modern flat UIs completely ignore our species intuitive ability to perceive depth.
Just yesterday I literally spent 20 seconds staring at the Word print dialog not sure how to actually make it print. There’s a flat image of a printer with ‘Print’ under it with a box around it but it’s at the top so I guess my brain wrote it off as a header image or something. I immediately lamented the lack of a clear ‘OK’ button, and felt almost crippling nostalgia for my first experiences of THREED.VBX back in the day.
For me, XP is the one I prefer. It looks more "modern" than 98 but not as "glossy" as 7. It has the right balance of better looking buttons and radios, but doesn't go too far
I love how the "Show code" buttons in 98.css and XP.css are just plain labels, with no indication that they are clickable until you hover over them, as opposed to the easily visible buttons the CSS actually provides x)
Interestingly, the 7.css fixes this by showing them as expander links instead.
I think one of the major "sins" of those interfaces (that is still pervasive today) is the inability to select text on most labels. Back then, Google did not exist, but being able to select and copy the text on e.g. an error message is something that has since become very useful, and that's still not afforded by GUIs.
Well, at least nowadays we can use ridiculous solutions such as normcap (taking a screenshot and OCRing it) to get such a basic feature.
On Windows, you can press Ctrl+C to copy the contents of native message boxes. Admittedly not very discoverable. They should have added it to the System menu as well.
A general facility to select labels should use an additional modifier key, since otherwise unsophisticated users will tend to select labels by mistake and get confused. Labels are already actionable elements, in that clicking on them focuses the associated control, or even changes its state as in the case of check boxes and radio buttons. Another example are the labels of menu items and buttons. Adding selection when you move the mouse a bit while hitting a label is too finicky IMO. If the UI is also a touch interface, it would only get worse.
From what this website says (https://weblogs.asp.net/chuckop/110153), this behavior started on Windows 2000, which might explain why it didn't work on my emulated Windows 98...
My solution to this is to use Emacs whenever I can. (Almost) any text in Emacs is in some buffer, so you can usually mark (select) it, kill (cut) it etc.
On that note, the css property user-select [0] may be useful to anyone trying to implement this kind of usability.
It can also be used to ensure a whole block of text is selected and copied as a unit. That is helpful for things like a an error message or an identifier like ABC-123.
I guess this is just a case of baby duck syndrome, which is so common in UI and UX discussions. I just don't see how word 97 is in any shape or form better than say, office 2019 in terms of UIs. Unless you're into obscure sub Sub menus in a single menu box that has no hints as what it contains or what it does. No preview, nothing.
We only very briefly had that around the start of Win95, by the time 98 came around (IIRC) Explorer and definitely IE already had flat toolbar buttons that only had bevels when hovering. I remember fetishizing how clean those new style controls looked (esp IE5ish with its grey-on-grey patterned background texture!) but it definitely broke the consistent interaction idiom you mention
In the case of toolbars though, I think flat buttons were OK given that the toolbar itself can easily be identified as a toolbar and you know that toolbars always contain buttons, so there is no need to identify each individual button as a button. Of course, when other widgets start getting added to the toolbar it's not as clear anymore.
with enough effort you can get win 10 (and maybe win 11) to look and function at least a bit like win98
I always stick with whatever pc I get at work until they require me to hand it back after so many years then I spend some headache to get it looking and working like a pc of yesteryear...
It wasn't really consistently enforced though. Just look at the intro screenshot and notice that every button in the toolbar is essentially just an icon with a subtitle. Basically just an uglier version of today's flat UIs
Mozilla SeaMonkey still has the HTML editor (and has been updated within a year!). SeaMonkey is strange because it looks like it was pulled straight out of 1999, but still has a Modern version of FireFox in there, even with the F12 web developer tools, despite also having the historical version of those same web developer tools there too.
I quite often need to update some of our static web pages. Not only does SeaMonkey do the job well, I'm not aware of anything else remotely similar that is still maintained (admittedly I haven't looked in a while).
Lightweight static HTML is a winner for many uses, and I'm honestly puzzled why there is so little of it remaining.
> Mozilla SeaMonkey still has the HTML editor (and has been updated within a
year!)
+1 to that. I still use SeaMonkey every time I need to edit HTML docs directly. Works great, haven't found anything better. (Admittedly haven't looked too hard for anything better because SeaMonkey works just fine.)
There's a local weightlifting gym near me that has a website line this. Checking the source it looks like it was produced with done obscure site builder and it's hosted on the owner's ISP hosting.
The html is simple but functional. The color choice is a bit garish.
But I love it. It loads quickly, it's always up to date and tells you exactly what you need to know. When is the gym open, where is the gym, how much does it cost. Done.
I feel we've lost this kind of experience on the modern web.
nah, i think that site needs to be a SPA of several megs of js so it pauses for a while on mobile before opening a modal welcoming me to their website with a full side bar of navigation links that fade in the new info before breaking and requiring refreshes for some sections of info to load because too many xhr
I learnt HTML purely because I was sick of waiting for Netscape Composer to launch on my 80386 PC back in the mid nineties. It had an 85MB WD HDD and 4MB of ram, running Windows 3.1. Double-clicking on the Netscape Composer icon in Program Manager would cause my hard drive to thrash for a solid 5 minutes, before the screen would update with _some_ resemblance of a Netscape window.
There's gotta be a better way to create web pages -- so I learnt HTML and was using notepad.exe to hand-craft my HTML pages. I would then copy them to disk (usually using ARJ.EXE to compress everything), and then go to school to use their internet connection to upload my pages to Geocities.
> Netscape Composer to launch on my 80386 PC back in the mid nineties. It had an 85MB WD HDD and 4MB of ram, running Windows 3.1.
Bleedin' Nora: that's certainly an optimistic system configuration for that application. I'm not surprised you lost patience with it.
I remember using a real mixed bag of tools to create HTML in the late 90s/early 2000s. Started off with Notepad but stayed away from Communicator until I had a PC powerful enough to run it easily (early y2k).
I also briefly tried exporting Word documents as HTML, which I think might have been new in Office 2000. This was a bad idea: the markup was hugely bloated, and images were primarily embedded as ActiveX objects that only looked good in IE, with heavily downscaled/coloured versions available for other browsers. Similar issues with Frontpage.
But I found the markup generated by Composer to be pretty clean by the standards of the time, so developed a hybrid workflow where I'd rough out pages, along with their content, in composer, and then tweak the markup manually.
I also remember finding a really nice text editor for working with web pages. It came free on a magazine cover CD and I wish I could remember what it was called [EDIT: it might have been HoTMetaL]. For editing raw HTML and JavaScript nothing could better it. It wasn't as good as VSCode + the right extensions today but, for the time, it was literally streets ahead. So I ended up using that + Composer for at least a couple of years, up until maybe 2002.
Word 97 has sensible export to basic HTML. No specific web page editing tools, but allows conversion of text content of existing documents in bulk. Search for
<META NAME="Generator" CONTENT="Microsoft Word 97">
Later versions of Office use HTML as some kind of complete document serialization format. The website
Thanks, I stand corrected. The export from Word 2000 was anything but basic. And it did, as far as possible, seem to try to serialize most aspects of the document. Which, of course, is not really what you want from a 1999/2000 web page that most people would view over dial-up.
I suspect their target market with this was enterprise intranets where everybody would be forced to use IE, and therefore all the ActiveX garbage would render just fine... and given LAN bandwidth most people probably wouldn't notice the ridiculously large payload sizes (for the era) of these pages.
I didn't know about the HTML filter though because I only experimented with the export once or twice, during the evenings after lectures, which was enough to convince me I was heading down a dead-end path.
That's basically the system I had in the mid nineties, though it had 8MB of RAM and was running OS/2 (2.1 to 3). I edited HTML on it in Emacs, mainly. Eight Megs And Continually Swapping was very accurate at that time.
I remember my mate's dad bought a 233MHz Pentium II system with 32MB of RAM some time in 1997, and our minds were all absolutely blown by the incomprehensible power of this system.
I want to say before that, back in 94, he'd had some sort of 486 variant (might have been a 486SX 33 or something along those lines - fairly run of the mill for mid-'94) with 4MB of RAM back in '94, and I want to say it got upgraded to 8MB, that we played a lot of DOOM and DOOM II on, and I think there'd maybe been a stop at a Pentium system somewhere in the middle because I remember we played Quake round at his place, and I'm sure that was earlier than '97. I'm not sure the 486 lasted that long - maybe only a year or 18 months - before it was replaced with a Pentium I system, but my memory is hazy.
I do remember the Pentium II though because, for the time, it was such a beast of a system. I do remember it seeming completely ridiculous and like you'd never need that much computing power. Oh sweet summer child, etc.
Sure, it was very ambitious to run something like a 16-bit version of Composer on such a low spec machine, but I still gave it a go :D
This was around 1995. I didn’t suffer too long with Composer on my 386, I got sick of the swapping pretty quickly and stuck with Notepad for a long time afterwards.
I didn’t upgrade my 386 PC until 1998 (couldn’t afford to), when I got a Cyrus 6x86, which is another story altogether. If you know about Quake and FPU performance with the 6x86, then you already know the story.
> I think the main thing I will remember about this experiment is: I had fun! It was a pleasant stroll down memory lane and a nice reminder of what HTML used to look like.
I literally still edit html files and upload them with ftp. What do people do these days? ahah
> discovering the "Insert Special Character" feature along the way (that I hoped would contain :smiley:)
Heh, back then, "special characters" meant "characters you get in languages that aren't English but still use a mostly Latin alphabet". The characters in the screenshot look pretty much like the high-bit-set portion of ISO-8859-1. If you wanted the Forrest Gump smiley face, then you had to use Wingdings.
Horizontal centring was never hard in CSS, it's just `text-align: center` for inline elements and `margin: 0 auto` for block elements. It's vertical centering that was impossible for many years.
margin: 0 auto adds an equal amount of margin to its left and right. This often does have the effect of getting the element centered, but may or may not be what you really need; e.g. it doesn't work if you need real margins, or if you want to nest aligned boxes.
For center tags, browsers to this day use a separate vendor-prefixed property, e.g. text-align: -moz-center. It is still much easier to use than whatever margin/float/flexbox/grid hack standard CSS has come up with, but -moz-center seems to be doomed to an eternal life as a non-standard (albeit ubiquitous) property.
(I remember that not so long ago even Google would set -moz-center on their search page. Such a great opportunity to standardize existing practice, left unexploited for 27 years and counting.)
Also, if anyone reads this, can you go to my site and tell me if the little pokemon minigame in the bottom right corner works for you? It usually works for me but sometimes if I open the site in Chrome on guest mode, it just loads a blank screen, but then if I resize the window it starts working. Sigh.
I might as well, but I find the <center> tag more charming. It's how my sister taught me to do it when I was learning NeoHTML for my Neopets pet page. After all, it's just for fun. In a different job market, I'd be worried about potential employers scrutinizing my code haha
Yeah, but that was less about centering in the web, and more about fonts in general tend to be made incorrectly which makes vertical centering hard everywhere, including the web.
There where several topics around centering (horizontal and vertical) and of course much more detailed than a simple "center"
But this [1] looks as something one would archive with a "center" back in the day.
It was meant for very basic pages as I found out the hard way. It would freeze up with even moderately complex table layouts. I used it for my first paid website build. With money made I bought an Iomega Zip Drive.
I wonder where all the stuff I used back then is today. The plastic and metal scrap of late 20th century e-waste. Somewhere buried deep.
I used Netscape Composer and then the similar component of SeaMonkey to maintain several static websites from around 2000 to 2015 or so. They were a bit buggy, but I was able to make them work well enough, and they produced much cleaner code than the other WYSIWYG editors I tried at the time.
I finally stopped using SeaMonkey when I felt the need to make my sites responsive. I now maintain my sites with a text editor and an FTP app, which work fine but feel like a step backward.
With my best friend, in our teenage we used to maintain our website using Netscape composer.
I'm very surprised that it came out this year because I believe we started our website maybe before 1997, but don't remember using another tool.
It was pretty easy to use for childs with zero help at all from adults. But was crashing pretty often (so where many apps in this era).
The hard part for childs was to host the website, hopefully, in France was had the free hosting association mygale.org
I'm so sad my bestie deleted our website with no backup without asking me about.
Those are fond memories that I'm pretty sure are the reason I because a developer.
I went to high school at the time (in Europe) and could create decent looking websites. I could have made a decent income from making web sites for businesses, but the idea of finding someone to host it on proper servers (with domain names) and actually sell and market it to customers were too foreign to me. Computers were expensive, so even if I found a place to host it for cheap, the machine was too much for my budget. You couldn't sell a website to a business and place it on Geocities...
Looking back at it now, the only truly difficult part (for me) was the infrastructure. I have heard stories about people just cold calling a merchant and sell a quick "home page" for $500 + hosting fees within minutes, and just generated a static page from a template with a logo, name and contact details. It was a gold rush where the right people with the right skills could make easy money.
Some of them were lucky enough to live in cities with local ISPs and hosted the sites on their personal PCs on a dial-up. People didn't expect much from the web at the time. I simply didn't have that kind of knowledge and risk taking ability at the time. My personal network did not contain any people who could have pointed me in the right direction either.
I'm kind of envious of today's youth, because they seem to have a better awareness of the business side of things and easily available technologies.
If my memory is right, FileZilla could resume interrupted downloads, keep downloads going on from multiple sites, store credentials, etc. Browsers had the functionality but it was more conveniently organized in FileZilla and similar tools. There were a zillion of download managers of all sorts, maybe still today but I'm not using them much anymore.
> The result pretty much occupies the whole viewport, with no margin. I did not find a way to add margins or padding to the page's body in Netscape Composer.
Should had just threw the contents in the table.
Also, every time some idiot assumes my 43" 4k display in the landscape mode is just a strange 6" smartphone and renders the site with gigantic text with 35% of viewport width - I just really want to punch them in the face over TCP/IP.
Please, don't be an asshole to your visitors, let them (and their devices) decide how to render the information you present to the whole world to see.
My first non-geocities website was built entirely in Netscape Composer. Eventually I got frustrated with it's limitations and started editing the HTML myself. I later briefly switched to Microsoft Frontpage before eventually settling into basically always running HTML by hand.
This is all documented in the news feed of the site that is still needlessly online.
Microsoft Frontpage is where it is at.
I wonder how many thousands website was created using Frontpage
by folks who in general had no knowledge about html/css.
I miss it.
The dirty secret of web apps is that we’ve mostly gone backwards in usability compared to native desktop apps from 25 years ago.
Web apps are a mishmash of paradigms. Pieces of desktop UI are reproduced using a woefully limited framework inside a static request-based page navigation model. The user never quite knows whether an action will trigger a multi-second page refresh, whether the back button does anything useful, etc.
Desktop UIs had professionally designed human interface guidelines based on decades of actual research. On the web, designers are primarily graphics artists who pick fonts and pride on making buttons look like nobody else’s buttons. Icons are nowadays tiny monochrome line scribbles without labels. Just pray there are tooltips so you can figure out what happens if you press one of these icon buttons in a web app. (Or maybe it’s just an icon and not a button? No way of knowing, since the conventions that made buttons obvious have been thrown away.)
The web is the worst application delivery platform of the past 30 years, so of course it’s the one we got stuck with. Worse often wins by its simplicity and ubiquity. Everybody could author a HTML page and some gradually built their skills towards apps. This review of old Netscape Composer reminds of how important that was.
Lazy, inaccurate, and offensive.
How is this limited to the web? 'The Desktop' suffers significantly more from not having a single, understandable paradigm. Is this UI modal? If I click this will it 'pop' me to another view? Or will it open a new window? Is that window represented in task bar/switcher/dock/whatever? How do I get back to where I was before?
Not to mention the issue that both macOS and Windows suffers from being in the middle of a UI-kit transition, with a bunch of half-and-half apps across the system, with the new system being of... dubious quality (I'm staring at you, new macOS Settings)
I think overall, I think there is a positive trend for usability of software on both the web and 'desktop' (whatever that means). But still, just like 20 years ago, people can still make bad software.
Edit: lol - just now as I'm trying to uninstall some AMD drivers, I get a dialogue box asking me to visit a long Microsoft Store URL, and it's neither clickable nor copyable. Desktop is perfect!
To take one example, 99% of monitors in 1997 were 1024x768 CRT with single pixel density, which freed designers to assume that's what their end user was on. No such standard exists today.
My bet it that, if computing doesn't evaporate in the next decade, there will be a middle-ground solution.
* Professional opinions drawn in higher ranked stakeholders whims * It drown visibility and attention to other processes of software development such as infrastructure, integrations, etc.
However, that also gives some nice things: Stakeholders and designers can fight in their own corner.
When I fully switched into the web dev world around Y2K, it felt so liberating. Never regretted it for the second...
This is nothing new. Amateurs going pro. DTP saw something similar, video as well.. lots of things.
What's new however is in mobile age, what are you even designing for? It's a moving target between device aspect ratio and screen sizes.
No wonder it's all a giant galleria abominate.
I wasn't aware it was a secret! Though it certainly isn't just web apps: the desktop has taken steps backwards too, and not just where bad web apps are hosted in simple desktop wrappers.
Back in the 90s and 00s I remember a strong push for consistency and discoverability, where now that often takes a back seat to aesthetic matters and shiny new things. And it isn't just the apps: whole OS UIs have fallen apart, with levels of inconsistency well beyond that which Mac and Windows people used to poke fun at the Unix world for. It is amazing how hard it can be sometimes to see at a glance which app currently has input focus on a Windows desktop, particularly with multiple screens.
There are several parts to the consistency problem, and the overall UX-is-more-shite-these-days problem, including:
1. The conflicting priorities of the range of devices (trying to make something work on a small touch-based mobile and a large desktop environment, often creating something that is imperfect on both).
2. The desire to be subtle and pretty, even where this makes things less obvious (some flat designs are particularly bad for this: little indication of what is actionable, scrollbars becoming single lines of pixels or invisible entirely, title bar text being slightly-whiter-than-mid-grey when active and slightly-darker-than-mid-grey otherwise and sometimes the reverse).
3. The desire to have an app that is easier to monetise where a plain web-app would be better (so you get something that is trying to optimise for four distinct UX arenas: web, mobile web, mobile app, desktop app).
4. The desire for everything to be online-first, so you can more easily justify charging for a subscription (and so you don't have to worry about sync issues inherent in offline-first design) even in cases where entirely offline would be the better choice for the user.
5. The desire to be hip and new with UI design completely overriding years of earlier UX research because form sells better than function. Yes, some of those research results may have been questionable, and some that weren't are less relevant to today's environments, but they all seem to have been thrown away in many places.
It isn't all bad though, some things are still well-designed from a UX standpoint, or if not well-designed at least not brain-dead. I live in hope that people will learn from the few good examples, and we'll have a swing back towards caring about properly optimising for UX instead of just looks & monetisation.
The audience of apps since smartphones came up (let's say 20 years ago) has vastly diversified, from previously mostly professionals and enthusiasts, to a far more general audience. That adds new requirements, and shifts priorities - it's more important now that simple things are simple, and it's less important that difficult things are possible at all. And from an inclusion pov, that's a good thing!
I'm not saying that everything has strictly improved, nothing ever does. But it certainly hasn't gotten strictly worse, nothing ever does either. It's just the tradeoffs that are different, and no longer target as specifically those folks who tend to yell at clouds on HN these days.
I’m also hard-pressed to come up with anything that has improved on the desktop.
Yes, yes I am.
Maybe it’ll work if I turn it off and on again.
I thought you were referring to the Ubiquiti web app.
It almost works in that sentence.
[1] https://jdan.github.io/98.css/
[2] https://botoxparty.github.io/XP.css/
[3] https://khang-nd.github.io/7.css/
Interestingly, the 7.css fixes this by showing them as expander links instead.
Well, at least nowadays we can use ridiculous solutions such as normcap (taking a screenshot and OCRing it) to get such a basic feature.
A general facility to select labels should use an additional modifier key, since otherwise unsophisticated users will tend to select labels by mistake and get confused. Labels are already actionable elements, in that clicking on them focuses the associated control, or even changes its state as in the case of check boxes and radio buttons. Another example are the labels of menu items and buttons. Adding selection when you move the mouse a bit while hitting a label is too finicky IMO. If the UI is also a touch interface, it would only get worse.
From what this website says (https://weblogs.asp.net/chuckop/110153), this behavior started on Windows 2000, which might explain why it didn't work on my emulated Windows 98...
It can also be used to ensure a whole block of text is selected and copied as a unit. That is helpful for things like a an error message or an identifier like ABC-123.
[0] https://developer.mozilla.org/en-US/docs/Web/CSS/user-select
First versions of OSX were fine too.
I was flabbergasted when I first saw the flat toolbars in MS Office, oh, whatever version that was.
And speaking of flat surfaces instead of buttons with depth: Laptop keyboards these days :-(
I always stick with whatever pc I get at work until they require me to hand it back after so many years then I spend some headache to get it looking and working like a pc of yesteryear...
If you want to test it out, make sure you install https://github.com/gorhill/uBlock-for-firefox-legacy/release... from there, because it won't show up in the extensions search.
And I use it regularly.
I quite often need to update some of our static web pages. Not only does SeaMonkey do the job well, I'm not aware of anything else remotely similar that is still maintained (admittedly I haven't looked in a while).
Lightweight static HTML is a winner for many uses, and I'm honestly puzzled why there is so little of it remaining.
+1 to that. I still use SeaMonkey every time I need to edit HTML docs directly. Works great, haven't found anything better. (Admittedly haven't looked too hard for anything better because SeaMonkey works just fine.)
The html is simple but functional. The color choice is a bit garish. But I love it. It loads quickly, it's always up to date and tells you exactly what you need to know. When is the gym open, where is the gym, how much does it cost. Done.
I feel we've lost this kind of experience on the modern web.
that would be way cooler than a usable website
There's gotta be a better way to create web pages -- so I learnt HTML and was using notepad.exe to hand-craft my HTML pages. I would then copy them to disk (usually using ARJ.EXE to compress everything), and then go to school to use their internet connection to upload my pages to Geocities.
Bleedin' Nora: that's certainly an optimistic system configuration for that application. I'm not surprised you lost patience with it.
I remember using a real mixed bag of tools to create HTML in the late 90s/early 2000s. Started off with Notepad but stayed away from Communicator until I had a PC powerful enough to run it easily (early y2k).
I also briefly tried exporting Word documents as HTML, which I think might have been new in Office 2000. This was a bad idea: the markup was hugely bloated, and images were primarily embedded as ActiveX objects that only looked good in IE, with heavily downscaled/coloured versions available for other browsers. Similar issues with Frontpage.
But I found the markup generated by Composer to be pretty clean by the standards of the time, so developed a hybrid workflow where I'd rough out pages, along with their content, in composer, and then tweak the markup manually.
I also remember finding a really nice text editor for working with web pages. It came free on a magazine cover CD and I wish I could remember what it was called [EDIT: it might have been HoTMetaL]. For editing raw HTML and JavaScript nothing could better it. It wasn't as good as VSCode + the right extensions today but, for the time, it was literally streets ahead. So I ended up using that + Composer for at least a couple of years, up until maybe 2002.
http://mc-computing.com/HTML_Examples/html_Generators.htm
reminds that there exist Microsoft's own clean-up tool (without doubt, an internal pet project which became essential), “Office 2000 HTML Filter”.
I suspect their target market with this was enterprise intranets where everybody would be forced to use IE, and therefore all the ActiveX garbage would render just fine... and given LAN bandwidth most people probably wouldn't notice the ridiculously large payload sizes (for the era) of these pages.
I didn't know about the HTML filter though because I only experimented with the export once or twice, during the evenings after lectures, which was enough to convince me I was heading down a dead-end path.
I want to say before that, back in 94, he'd had some sort of 486 variant (might have been a 486SX 33 or something along those lines - fairly run of the mill for mid-'94) with 4MB of RAM back in '94, and I want to say it got upgraded to 8MB, that we played a lot of DOOM and DOOM II on, and I think there'd maybe been a stop at a Pentium system somewhere in the middle because I remember we played Quake round at his place, and I'm sure that was earlier than '97. I'm not sure the 486 lasted that long - maybe only a year or 18 months - before it was replaced with a Pentium I system, but my memory is hazy.
I do remember the Pentium II though because, for the time, it was such a beast of a system. I do remember it seeming completely ridiculous and like you'd never need that much computing power. Oh sweet summer child, etc.
This was around 1995. I didn’t suffer too long with Composer on my 386, I got sick of the swapping pretty quickly and stuck with Notepad for a long time afterwards.
I didn’t upgrade my 386 PC until 1998 (couldn’t afford to), when I got a Cyrus 6x86, which is another story altogether. If you know about Quake and FPU performance with the 6x86, then you already know the story.
> I did not find a way to add margins or padding to the page's body in Netscape Composer.
Full page width 3-column table, with your main content in the middle column?
I literally still edit html files and upload them with ftp. What do people do these days? ahah
Heh, back then, "special characters" meant "characters you get in languages that aren't English but still use a mostly Latin alphabet". The characters in the screenshot look pretty much like the high-bit-set portion of ISO-8859-1. If you wanted the Forrest Gump smiley face, then you had to use Wingdings.
margin: 0 auto adds an equal amount of margin to its left and right. This often does have the effect of getting the element centered, but may or may not be what you really need; e.g. it doesn't work if you need real margins, or if you want to nest aligned boxes.
For center tags, browsers to this day use a separate vendor-prefixed property, e.g. text-align: -moz-center. It is still much easier to use than whatever margin/float/flexbox/grid hack standard CSS has come up with, but -moz-center seems to be doomed to an eternal life as a non-standard (albeit ubiquitous) property.
(I remember that not so long ago even Google would set -moz-center on their search page. Such a great opportunity to standardize existing practice, left unexploited for 27 years and counting.)
https://tonsky.me/blog/centering/
[1] https://tonsky.me/blog/centering/apple_sign_in_business@2x.w...
I wonder where all the stuff I used back then is today. The plastic and metal scrap of late 20th century e-waste. Somewhere buried deep.
I finally stopped using SeaMonkey when I felt the need to make my sites responsive. I now maintain my sites with a text editor and an FTP app, which work fine but feel like a step backward.
It was pretty easy to use for childs with zero help at all from adults. But was crashing pretty often (so where many apps in this era).
The hard part for childs was to host the website, hopefully, in France was had the free hosting association mygale.org
I'm so sad my bestie deleted our website with no backup without asking me about.
Those are fond memories that I'm pretty sure are the reason I because a developer.
Looking back at it now, the only truly difficult part (for me) was the infrastructure. I have heard stories about people just cold calling a merchant and sell a quick "home page" for $500 + hosting fees within minutes, and just generated a static page from a template with a logo, name and contact details. It was a gold rush where the right people with the right skills could make easy money.
Some of them were lucky enough to live in cities with local ISPs and hosted the sites on their personal PCs on a dial-up. People didn't expect much from the web at the time. I simply didn't have that kind of knowledge and risk taking ability at the time. My personal network did not contain any people who could have pointed me in the right direction either.
I'm kind of envious of today's youth, because they seem to have a better awareness of the business side of things and easily available technologies.
And I thought everyone knew smileys weren't part of Unicode until about a decade later.
[0] https://en.m.wikipedia.org/wiki/GetRight
Today the BlueGriffon Editor[1] fulfills a similar purpose. At least that is what my daughter used in class.
[1] https://en.m.wikipedia.org/wiki/BlueGriffon
I wish there was a WYSIWYG editor that has CSS support and modern HTML.
Should had just threw the contents in the table.
Also, every time some idiot assumes my 43" 4k display in the landscape mode is just a strange 6" smartphone and renders the site with gigantic text with 35% of viewport width - I just really want to punch them in the face over TCP/IP.
Please, don't be an asshole to your visitors, let them (and their devices) decide how to render the information you present to the whole world to see.
This is all documented in the news feed of the site that is still needlessly online.
it works. always.
https://developer.mozilla.org/en-US/docs/Web/HTML/Element/ce...