Am I understanding the setup right?
And re-watch this also - Project Code Rush - The Beginnings of Netscape / Mozilla Documentary https://www.youtube.com/watch?v=4Q7FTjhvZ7Y
Is this normal for wiki pages on people?
One day I read about a guy in brooklyn who had a website at www.soundtube.com and was selling music on the internet . I got in touch and went to his office in brooklyn to look at his website in a graphical browser. I than followed his lead in getting setup.
The logo for the site was a half squeezed tube of toothpaste with the word sound tube on it.
I don’t remember his delivery mechanism. The last time I visited the site it was the same logo but with the subtext that “what could have been”.
I occasionally look for more information about sound tube.
Seems to be lost but I hope it is only missing.
> "We've got to make progress on [renaming the company]." And I said,
> "We've got a couple of ideas, but they're not great." Then it just kind
> of popped into my head, and I said, "How about Netscape?" Everyone kind
> of looked around, saying, "Hey, that's pretty good. That's better than
> these other things." It gave a sense of trying to visualize the Net and
> of being able to view what's out there.
Greg Sands in https://money.cnn.com/magazines/fortune/fortune_archive/2005...A few days ago JWZ had a great take on where Mozilla is today: https://www.jwz.org/blog/2024/06/mozillas-original-sin/
No lack of truth or taste in punctuation is implied by this edit.
http://home.mcom.com/mozilla.org/1998-03-25/party/
That party was a huge milestone in retrospect. It was the day FOSS went mainstream. Shortly thereafter, the dot-com boom ended and the 90s tech parameters got upended and scrambled.
Of course this was Mosaic. And of course I was totally and completely wrong. Said he while using the Firefox web browser. And when was the last time I used telnet?
I'd always choose Internet Explorer because of this. I'm really glad that Netscape rebranded to Mozilla Firefox. Much warmer and more inviting, less implied threat of drowning.
And what I love the most about these guys (Marc, PG, even Sam Altman) is that they ARE hackers. They speak in our terms, they have our awkwardness.
Thanks for sharing this.
At some point he did write why he bought the club, he was moaning about the state of night life in SF, and a friend said something like "Why don't you do something about it?"... so he did.
Edit: found it: https://www.dnalounge.com/backstage/log/1998-1999.html
However, since I was the Project Lead for the Spyglass browser team, there is one correction I can offer: We licensed the Mosaic code, but we never used any of it. Spyglass Mosaic was written from scratch.
In big picture terms, Marc's recollections look essentially correct, and he even shared a couple of credible-looking tidbits that I didn't know.
It was a crazy time. Netscape beat us, but I remember my boss observing that we beat everyone who didn't outspend us by a favor of five. I didn't get mega-rich or mega-famous like Marc (deservedly) did, but I learned a lot, and I remain thankful to have been involved in the story.
The hierarchy there was basically a reflection of the company's browser team org chart. You could find a group for every team working on the browser where many of them were having their regular technical conversations.
I remember reading “you can go to the Louvre and then the MoMA, all with a click of the mouse”. But taking a plane felt almost as slow and expensive, only way more fun.
I deleted Netscape to claim back the 20MB or so it occupied in my 250MB drive.
While Marc recounts that Microsoft offered for Spyglass to sell "Microsoft Mosaic" as an add-on while still offering your own independent version - despite MSFT eventually making its own browser free anyway - is there anything within that part of the larger story that you would elucidate to tell differently, or clarify deeper into its weeds? It was always one of the parts of the story that was more glossed over.
I’ll try to go to news.ycombinator.com and Lynx tries to make an NNTP connection and I don’t blame it.
I can’t recall the exact timing of when NCSA ceded all sublicensing rights to Spyglass. It may have been after that experience or a relief that they could send MS away in good conscience.
It’s hard to compete with free. And the NSF asked several times if they should still be funding it.
The following summer I applied to work there. I did not miss the next several shifts in the market, but eventually got tired of chasing them.
IIRC, he decided a long time ago that he'd had enough of crazy startup life, and bought a nightclub, and somehow kept a nightclub going all that time.
That was a very complicated day.
The next summer, I was at the University of Florida, but off-campus. However, the Alachua [County] Freenet offered free dialup with PPP. Since etherppp emulated an Ethernet packet driver, the NCSA apps worked fine there, though obviously much slower.
Better, more complete DOS-compatible suites have arisen since then (e.g., mTCP), but the NCSA suite was fantastic. Security? Nah, none of that. But useful? OMG yes.
It was spending the fucking money.
Foundations like some cancer groups and the arts have an endowment. Each year they build up their war chest by seeking new funding, but a lot of the money they spend each year is the interest payments on their giant piles of cash. Mozilla could have run in perpetuity on the money Google gave them, but instead they decided to branch out into boondoggles and dipping their hands into the cookie jar.
I mean. It’s still very far from actually being NNTP, and it’s not decentralized like Usenet or anything like that.
But all this time I’ve been thinking of Slack as “better IRC, with images and links and threads”.
When really Slack is more like “fancy Usenet service with client that renders images and other attachments”. (Although on the protocol and server and client implementation level it is very different from NNTP.)
Well. At least we don’t have to inefficiently yEnc encode attachments nor to split attachments into a bunch of pieces with par2 files. So there’s that.
I could have gotten in on the third big round of hiring at Amazon, but I told my friend I’d rather work until retirement than get rich writing Perl code. People are allowed to have standards, and those standards are allowed to keep you from taking money you don’t feel good about.
If it wasn’t then we would all be sex workers. Most pay for the least work.
In my recollection, the initial payment from Microsoft to Spyglass was higher than what Marc said, but I'm not sure.
But I am sure that the deal was later renegotiated at a substantially higher number.
I'm also pretty sure that even after that rework of the terms, Spyglass didn't get enough from Microsoft to compensate for the fact that Microsoft, er, you know, killed the browser business. And insofar as that is the essence of Marc's point, I agree with it.
"The Microsoft guys call Spyglass and they're like, yeah, we want to license Spyglass Mosaic so we can build it into Windows. The Spyglass guys say, yeah, that sounds great. Basically, how much per copy are you going to pay us for that? Microsoft says, you don't understand, we're going to pay you a flat fee, which is the same thing that Microsoft did when they originally licensed DOS way back when. But Microsoft said, basically, or at least my understanding of what Microsoft said was, don't worry about it. We're going to sell it as an add-on to Windows. We'll have Microsoft Mosaic and then you'll still have Spyglass Mosaic and you can sell it on other operating systems or compete with us or whatever, do whatever you want."
Thank you for your response!
Why people might want to adjust the `Referer` behavior of the browser is that it leaks more information than you might think.
As for his personality, I get the impression he was always like that.
https://en.wikipedia.org/wiki/HTTP_referer
Etymology
The misspelling of referrer was introduced in the original proposal by computer scientist Phillip Hallam-Baker to incorporate the "Referer" header field into the HTTP specification.[7][8] The misspelling was set in stone by the time (May 1996) of its incorporation into the Request for Comments standards document RFC 1945[9] (which 'reflects common usage of the protocol referred to as "HTTP/1.0"' at that time); document co-author Roy Fielding remarked in March 1995 that "neither one (referer or referrer) is understood by" the standard Unix spell checker of the period.[10] "Referer" has since become a widely used spelling in the industry when discussing HTTP referrers; usage of the misspelling is not universal, though, as the correct spelling "referrer" is used in some web specifications such as the Referrer-Policy HTTP header or the Document Object Model.[3]
It was called Sound Wire, not Sound Tube - which is probably why you couldn't find anything... perhaps the name got mixed up with the toothpaste logo in your memory. Memory does that!
https://web.archive.org/web/19961122055147/http://soundwire....
https://www.wired.com/1995/05/net-surf-44/
p.s. I messaged him - maybe he'll show up in the thread
On the other hand, the Google money accounted for around 85% of their income over the years, so if they hadn't been spending it they would have been operating on around 20% of the income for many years while the endowment grew, and likely would not have been able to keep up with competing browsers.
Also, for as much crap as she gets, Mitchell Baker invested over 20% of the Google money Mozilla received during her tenure, far more than was invested by prior CEOs. And before anyone brings it up, all that "woke activist" spending comes from donations, not Google money, which the IRS prohibits them from spending on browser development.
Then I installed Mosaic on my PC, and ran the Info documents through a converter to produce html. I showed my boss the documents with Mosaic, and this time he said Wow!
True. I created an online product catalog thing. For reasons I can't remember, I used SuiteSpot and JRunner.
Turrible. Absolutely turrible. Truly unforgivably bad.
Ditto their LDAP thing.
And Netscape sabotaged Java and Applets. And created JavaScript. And XUL. And...
But hey, marca famously named the image tag "img". So it wasn't all bad.
I ended up going to grad school instead of jumping on the gravy train. Still kicking myself for that to this day :-)
There is a huge overlap from groups I hung out with in high school and college (UCSC) and people that were at Netscape. There were a lot of super talented people.
There's something really special about the community and openness of folks who work on web browsers. Maybe it traces it's way back to the newsgroups.
At some point we came across and downloaded BBS lists like Focke's and software like Telix, and realized we didn't need to pay $9.95/mo for access to interesting communities. The local BBS's were way more interesting and niche (and longtail) than anything found on the moderated Prodigy anyway. The pressure of not pissing off "mom" for spending extra time on Prodigy, which had a pay-by-the-minute, access plan at the time was extra appealing even if we could only spend 30-45 minutes on a local board at a time. It was all so reasonable.
But local boards were ANSI and later ASCII and the graphics on Prodigy [2][3] were sorely missed -- which were about the equal of even the best EGA graphics of the time. Games were descriptive instead of graphical. But the local communities (who you could meet up with), the forums, and the price (free) were an appealing draw to an early teen with no money. RIP Graphics BBSs eventually arrived a couple years later but they were few, fussy, and were more representative of the (by then) aging Prodigy graphics than the new VGA and high-res Windows 3.x GUIs we were growing used to.
We had a buddy, the next town over, who was a major Apple Macintosh enthusiast. As a result, he generally eschewed the gross and primitive ASCII scene, but was as cash strapped as we were. IIR RIP BBSs sort of bypassed Macs, but a bizarre sort of Galapagos technology appeared in the form of full GUI BBSs. I remember one client called "FirstClass" [4] that basically just extended the resource of the BBS onto the Mac desktop. It was absolutely mindblowing, and included a primitive ability to request simultaneous data streams allowing you to view a forum and download an image or a file at the same time. There wasn't a good MS-DOS/Windows client so we spent hours and hours and hours at that friend's house blowing up their long-distance bill dialing in to any first-class number we could come across.
As a parallel track, in the early 90s, (maybe '91 or '92) my Mac buddy ended up with access to a dial-up Unix shell through their parents, who had it for work. We memorized the password and ended up freaking out as we learned how to gopher, ftp, and telnet to sites all over the world. The semantic binding of protocols://servicestypes made an astonishing kind of sense.
I found out about the demoscene around this time on dial-up BBSs, but I found the actual demoscene on open access anonymous ftp sites in Florida and Finland and other places around the world. The amazing movie Sneakers came out about that time and it dropped into our developing digital milieu like warm socks out of a hot clothes dryer on a winter day. My friend's father eventually discovered our account usage (because we were blowing up his corporate account bill), and we were locked out. But I knew at that point, that BBSs were now the second tier in the information landscape. Cyberpunk novels entered my life and I knew the internet = cyberspace, not BBSs.
I ended up in a special program through my school district that happened to include access to my own gopher/shell dial-up through the district. I had a luxurious 20 minutes a day and 1 or 2MB of storage to play with. But as a high-schooler, getting access to what I had only known as the realm of top universities or global corporations was thrilling. I learned how to exit the default gopher menu and use the other unix tools to ftp, telnet, and do everything else I needed to connect to what I inferred as other digital pioneers around the world.
I graduated in '95, lost my access to the internet, which felt like the loss of a limb and spent a a year relegated to the local BBS scene, which was still going strong. RIP had stalled, and the Mac gui BBSs were only a distant ideal of what could be. Modems were 14.4 or 28.8 baud.
I found out that some other friends were starting an ISP through some miracle, and I secured a job with them, quit everything else, immediately transitioned to living off of a T-1 8+ hours a day. I carried a hard drive in to work with me, connected it to a spare IDE port in my day-to-day desktop, downloaded what I wanted, and brought it home...like it was a thumb drive. It was a drug. BBSs died for me at that point -- I just...stopped dialing in to them. Very quickly we adopted this software called Mosaic, tied to yet another semantically aligned protocol called HTTP. It just slotted in the mix of telnet, ftp, nntp, smtp, gopher, and others. It was cool, but it took forever to load a page vs a gopher site or a telnet site. Usenet was the vibrant global forum that was the "big-boy" version of the local BBSs I had been using. I remember when Amazon first put up their website and sold only books. I didn't trust sending my credit card over the internet, so I'd find out about new books then go to local bookstores to buy them. For a year, I lived in the future.
At some point we decided to distribute Mosaic, then quickly after than I remember an early Netscape to new signups (along with dial-up sofware, email software, and Usenet software) -- the entire kit fit on two 1.44MB floppies, a version for Windows and Macs (copied by my old Mac First-class BBS buddy). The rest of the semantic protocol internet, other than email died then -- even if we weren't quite aware of it. Gopher became a ghost, ftp lived a while longer, telnet sort of existed, Usenet was a constant "should we still mirror it" question. We would have killed the rest except the dial-up software, email client, and Mosaic needed slightly more than 2 floppies, so we filled the rest of the second disk with more software.
Modems at 28.8 became normal, and we started get requests for 56k and ISDN.
I started using my access in the ISP to create unlimited time dial-up accounts for my friends. Girls I like dated me because I got them internet access, and members of the U.S. Demoscene suddenly could talk to their peers in Europe because of it.
Mosaic drove up bandwidth demand to astronomical levels. It was the Macintosh first-class BBS software realized to the nth degree. We move the ISP to the same building as our tier n-1 provider, drilled a hole in the concrete between floors and got rid of the t-1 by
We dropped usenet, ftp, and telnet clients off the disks. Dial-up software + email + Mosaic became the norm. ISDN turned out to be kind of a bust, DSL was on the horizon and we saw that it was the end of the mom-n-pop ISP because of how the technology worked. We sold the ISP and moved on elsewhere -- but Mosaic + email + dial-up became "the internet" from that point forward.
To be honest, I'm kind of sad to see PROTOCOL-OVER-HTTP came to erase the other semantic protocols. The way in which the browser kind of erased the rest of the internet has caused later generation from forgetting what could be possible over the internet. There's no reason at all that somebody can't come up with an entirely new protocol for a specialized service -- but the entire industry is stuck trying to figure out how to shove a square protocol into a circular HTTP(s) hole. This has allowed browser makers to really centralize and control large portions of the internet. It's like being told you must stick to specific roads when you are standing in the middle of an easily traversable, open, recently mowed, field.
If there is one thing I could will back into existence from OG internet is that concept. The Web IS NOT the internet.
1 - https://youtu.be/FNxKg6ZXax8
2 - https://www.theatlantic.com/technology/archive/2014/07/where...
I first used lynx years later when I was getting into Linux in the late 90s, and I found that part surprising at the time.
IMHO, it puts to shame the bloated, non-portable, overly-complicated, advertising-sponsored crap that is distrubuted today.
https://www.w3.org/Library/Distribution/w3c-libwww-5.4.2.tgz
30 small example programs written in C plus documentation for every one. Good luck finding something like that today.
I'm sure most younger people think of the internet either as the web (i.e. web pages you can access in your browser) or depending on age maybe just social media apps like TikTok and Snapchat, but of course the internet is just the network itself that connects everyone together, and then there are layers of software protocols (starting with TCP/IP) that support various apps on top of that.
If you're young the only protocol you may have heard of is HTTP (Hyper-Text Transport Protocol) which is what the web (World Wide Web) uses to send web pages from server to client (browser), which you are reminded of in web based URLs starting with http://www., where the www is also a reminder of the original "World Wide Web" name.
Other internet applications use their own transport protocols on top of TCP/IP to communicate, so we also have NNTP (Network News Transport Protocol) for UseNet, SMTP (Simple Mail Transport Protocol) for e-mail, and FTP (File Transport Protocol) for file transfer.
The power of the standard protocols was that they decoupled application from communications so that many alternate web browsers, e-mail clients, etc could exist and all happily communicate with servers supporting these protocols. A good example of what happens when you don't do this is instant messaging where originally the IRC (Internet Relay Chat) protocol was used as a standard, but later chat became balkanized into competing non-standard applications such as AIM, MSN and ICQ which were not able to inter-communicate until many eventually supported ICQ's Jabber/XMPP protocol. Even today instant messaging suffers from balkanization with iPhone and Android not able to share all features (blue vs green messages), although that is finally improving.
Nowadays most people have switched to web-based mail rather than using SMTP clients, but happily the e-mail servers still use SMTP to inter-communicate, so we can still send e-mail to each other!
The latest internet trend is all the social media apps - Twitter, TikTok, Snapshat, etc - which just like the instant messagers use their own proprietary protocols to talk to their servers, and are therefore not able to inter-communicate.
My reading is that jwz thinks there was a possible future where DRM is dropped because it's as useless & impractical to enforce as cryptography export restrictions had been. Mozilla could have contributed to this future by not implementing DRM, but instead supported the outcome we got: DRM is ubiquitous, browsers that don't support it are disadvantaged significantly, and an anti-DRM streaming service (similar to GOG) no longer has any real advantage over DRM-enabled services.
It is possible that no DRM in Mozilla would have resulted in the same outcome we arrived at - Mozilla gave in, so we'll never know. But what does Mozilla even exist for if it's unwilling to stick to its principles?
I wrote an online hypertext system in 1985, but the storage format was optimized to make it as efficient to transfer and display as possible and was not easy to author. It ran on top of the GEM GUI and you could click on a word that had been defined as a link to take you to the target page.
Someone could also have defined a rich-text schema in ASN.1 in the late 80s then written an application to retrieve data in this format from a remote server over an OSI network and display it. Interfacing the typical public text database of the time to this would have been a lot of work, they just expected to output to a terminal.
I was hired in Sept 1998 to work on the browser; we had built our own Javascript engine to add to it (since that was kinda required for the web by then). I rewrote all the table code, because it just really didn't work well when you had "too few" horizontal pixels, especially if table widths were expressed in things like %. In the end, after a major redesign of all the table code, it did better than Netscape did in the 'hard' cases.
However, before long, it became apparent with all the additions being made as part of HTML4 that sticking with Spyglass-derived code and trying to update it ourselves to compatibly implement HTML4 (or enough of it) was going to be a herculean effort for a small company (max ~350 people and briefly a $1B valuation (1999), but only around 5 or 10 people max on the browser, including the JS engine.
Given that, I made the decision in late 1999/early 2000 to switch us to the upcoming Mozilla open-source browser, and got deeply involved. The Internet-over-cable-TV part of the company failed (cable companies had other priorities, like breaking TVGuide's patent monopoly, which they paid us to do for them), and we moved onto other markets (hardware videophones) not involving browsers in 2003. I stayed involved peripherally in Mozilla, and when WorldGate dissolved in 2011 I joined Mozilla fulltime to lead the WebRTC effort.
The Spyglass internal architecture seemed at the time to be pretty reasonable compared to what I knew of the NCSA code.
https://web.archive.org/web/20010430044121/http://www.jeremi...
Posted it to slashdot at the time too, I miss those green colors ;)
https://slashdot.org/story/98/10/28/1923205/original-netscap...
If Mozilla refused to implement DRM in Firefox, Netflix would have just said “you need Silverlight, Chrome, or the native Netflix app to watch movies”, plain and simple.
I don’t feel bad because I would’ve sold it at $20 or $100 for beer money.
It's not like FF is a major browser that needs DRM to compete against Edge/Chrome. Its market share is in the single digits regardless.
So in that alternative universe we would likely have a non-responsive rectangle kind of UI that has to be loaded upfront. Despite all its shortcomings I much prefer the web, thank you very much.
Simply granting access to the host browser's DOM and event loop was intractably, technically, logically, morally, plantatively, confectionately, legally, politically, and in all other ways utterly impossible.
> non-responsive rectangle kind of UI that has to be loaded upfront
You probably never heard of the Shockwave/Flash runtime. Like Java Applets, Flash was also stillborn, for similar reasons.
Ha.
That said, Netscape's brain dead thread implementation, and seemingly unwillingness to even try to fix it, is what borked Java Applets. The success of the JavaScript, and now WASM, VM & runtime is proof enough of the feasibility.
##
Mea culpa: Upthread, I lied by omission.
Speaking of GUIs, you reminded me of Netscape's awesome Internet Foundation Classes. It greatly informed the subsequent Java Swing. https://en.wikipedia.org/wiki/Internet_Foundation_Classes
Architectually, IFC is Qt, more or less, but written in Java. As the successor to IFC, Swing is also awesome.
With the benefit of hindsight, Swing's embrace of MVC was an error. But "we" didn't know that then. (Design Patterns, amirite? It was phase. Sure seemed smart at the time.)
Ditto direct access to components vs requiring all state changes to go thru the event loop.
I can't defend AWT. I'm sure they had their reasons. Probably peer pressure. (Edit: An unintentional pun! Peer! Like AWT's component peers. Get it?! Gasp; I slay me.)
EVERYONE was so certain components had to be native and owner-drawn. Of all the prior cross platform GUI frameworks, AWT was simply the most ambitious, and so therefore the easiest to criticize.
Having previously shipped a few cross platform products, with the scars to prove it, I was completely against the strategy.
As someone who had used Prodigy since 1990 and began my CS program in fall of '94, I was extremely excited to get a much faster than 14.4k connection in my dorm room so I could use Mosaic at a proper speed. I seem to recall this was a universal feeling amongst other students in my class... like a "holy sh*t, I can't believe I have this level of access now".
But I only really started to enjoy the Web with 33.6k and by 56k I was completely hooked. That’s my recollection at least.
I loved IRC and BBS, but the Web took a while. Mostly because it was so slow.
If DRM weren't added to Mozilla and Firefox, then they would have continued to languish in marketshare on Windows/Mac and only would have hurt open source users on Linux/FreeBSD/etc.
The long-term gains of Firefox gaining marketshare (shaking up the IE monopoly and allowing web technologies to break stagnation) were worth the short term loss of "principals" on DRM. At least, IMO.
I had a similar experience with BBSs and I remember the romance of using a modem. I actually grew up in Champaign and recall dialing into “prairienet.” I also remember using AOL, but mostly to find and download games.
Netscape made a web user out of ppl where mosaic couldn’t.
Each contributed their progress.
Everyone had a braindead thread implementation at that time. Even Sun had to fix things in Solaris so that Java would run happily.
With the possible exception of Windows, you can than the Java Hype Cycle if you had a non shitty thread implementation in your operating system in the late 90’s.
NCSA went with Win32s to get threads and 32 bit addressing. I’m not sure what Netscape did, but by skipping Win32s it allowed them to ship on a single floppy for a long time. Which people found to be a feature.
NSCA has special dispensation to redistribute Win32s, so if you ordered a free copy of Mosaic you could get a copy of Win32s on another floppy.
- Lynx is far older than I thought. - WorldWideWeb 1.0 understood images, but didn't inline them, which is really what my creaky memory meant when it thought images were there from the beginning.
Improbably, because I had no idea which way was up, I did some contract work for Adobe bridging AutoCAD with Curo (a doc mgmt something). Win16 to Win32. Or vice-versa, I can't remember. "Thunking." An apt name. Seems like it shouldn't have even worked.
> thank the Java Hype Cycle if you had a non shitty thread implementation
I'm sure. I helped "Alligator Descartes" with the Magician OpenGL bindings for Java, around the time of Java3D. After much fussing, I managed to reliably get multiple contexts to render in sync. On Windows at least. Thank god for Doug Lea's work. Watching a bunch of cubes spin, each in their own window, was pretty neat for the time.
I watched this "Birth of the Browser" episode. It was a nice stroll down memory lane. Plus I dug a bit more to find corroborating stories. marca's recall of history and arcana is very impressive.
My low esteem for marca's acumen, abilities, and accomplishments was very wrong. Most importantly, his instincts and advocacy wrt HTML, HTTP, etc. was right, whereas I was mosdef wrong. (Repeatedly.)