Jump to content

Running Windows 98 in 2020 and beyond...


Wunderbar98

Recommended Posts

Agreed, believe we are Ying and Yang @siria. You are quite knowledgable and have been very helpful, thank-you very much :)

Regarding Windows ME Scan Disk, haven't noticed an appreciable difference but i only have a small drive and rarely perform surface scans. Query whether MS made error finding, speed or stability improvements. Just speculating though. I may not re-install ME's Scan Disk on a fresh system. Maybe someone else can comment if the ME version is actually any better.

Windows ME Disk Defrag is awesome, however. Extremely fast, mesmerizing compared to stock Windows 98. Microsoft appears to have done a great job at opening up a performance bottleneck. If uncertain, just install them after a full system backup, run Scan Disk and Defrag, see what you think. So far neither of these have caused issues, such as file loss, or appreciable instabilities.

I can not comment on KernelEx issues, other than those are the types of situations i have thankfully been able to avoid. Fortunately i have not yet found a need for newer software versions that could not be run in vanilla.

Although some may disagree and be perfectly content working offline, couldn't agree more @Bruninho. That's why connectivity and browsing has consumed so much of this thread. This system was originally intended to be offline for 'gaming'. Wasn't long before i wanted internet access to learn more DOS, perform research, download gaming-related files, check on news and weather, etc.

In case some readers have never used Windows 98 SE or just want to walk down memory lane, stock screenshots:
https://guidebookgallery.org/screenshots/win98se

Link to comment
Share on other sites


The oldest available NoScript extension was found to work in RetroZilla v2.2. It is good for anyone wanting fine-grained JavaScript blocking plus other stuff for safer browsing. Also good for older hardware that gets overwhelmed when loading JavaScript heavy websites. It should work in Firefox v2 and SeaMonkey v1.1.19 but has not been tested. Used only briefly, will update post if problems observed.

Go to https://noscript.net/getit. Download the oldest available NoScript version 2/3 down the page, right-click and select Save Link Target on 'latest NoScript version compatible with Gecko < 1.9 is 1.10'.

The usual manual install method did not work in RetroZilla:
File dropdown -> Open File -> select noscript-1.10.xpi file

A minor install.rdf file modification was needed:
- Unzip noscript-1.10.xpi file
- Open install.rdf in WordPad
- RetroZilla identifies as SeaMonkey
- Under !-- SeaMonkey -- reduce minimum version to 0.1
- Rezip the extension into *.zip format using exact same directory structure
- Rename noscript-1.10.zip to noscript-1.10.xpi
- Install noscript-1.10.xpi into browser

IMO the default whitelist allows too much through, remove items as desired.

Edit1: Recommend backing up the user profile before NoScript install, as it does not install in the usual manner (solely in to the profile's chrome sub-directory). If NoScript removal is desired either restore the backed up profile or manually delete the following from the profile directory:
- Entire 'ABE' sub-directory
- Entire 'pref' sub-directory if it only contains noscript.js, otherwise just noscript.js
- NoScriptSTS.db file

Edit2: Modified noscript-1.10.xpi file attached.

noscript-1.10.xpi

Edited by Wunderbar98
Link to comment
Share on other sites

I am a web developer, and I wish more sites could do what CNN does, they have a low-bandwidth version of their site. There was a trend of text-only websites back in 2017-2018. It was primarily done for critical situations, but it can also help us through our retro browsing activities.

You can read about them here and here:

https://www.poynter.org/tech-tools/2017/text-only-news-sites-are-slowly-making-a-comeback-heres-why/

https://www.maketecheasier.com/access-low-bandwidth-websites/

In case you'd prefer to go straight to the low bandwidth versions of some sites:

https://lite.cnn.io/en

http://text.npr.org/

Twitter and Facebook had one as well, but they $#*%$*%$ ported the experience to mobile apps only.

EDIT: Adding some more that I haven't tested yet:

https://en.m.wikipedia.org/wiki/Main_Page

 

Edited by Bruninho
Link to comment
Share on other sites

The quickest way I found to edit an xpi (or any other zip file), just not sure if it works with very old 7z-versions too??
HowTo:
- Open the xpi or jar or zip etc with 7-zip (not extracting, just opening in 7zFM filemanager)
- Open the wanted file by doubleclick (mine opens in N++), or if that doesn't work, by right-click/EDIT
- IMPORTANT!! After editing do NOT SAVE, just close the Editor by the X in upper right WINDOW corner (not just 1 tab, but ALL tabs at once)
=> A popup will ask if the edits shall be saved inside the xpi
=> Close xpi by Window-X
=> Another popup will ask if the xpi shall be saved

Took me years to figure out why direct editing worked only sometimes but more often not, and still not sure if that's now really the 100% final method ;-)

Link to comment
Share on other sites

One more method of editing for those who already have the required setup: in Total Commander if compatible packers are installed one can press Ctrl+PageDown on the selected xpi to open it as a folder, then navigate to desired file, click Edit (or press F4), do your changes, save the file in the editor, close the editor and accept the repacking. Very similar to the 7-zip method above.

Link to comment
Share on other sites

Thanks for all responses. All text and mobile links above worked in RetroZilla v2.2 without JavaScript - nice. Wish all developers created more accessible sites. When a banking site revamps it usually means more JS usage tracking and features that add little value. Amazing how many CPU cycles are now needed just to log in and check account details. Sometimes i complain but they need to reach a threshold before being noticed. Developer's ears close when they learn i am not using 'approved' software, all of my OSs and browsers.

For me, aside from incorrect code changes, is ensuring modified *.xpi extensions maintain the same directory structure. After modifying an extension, zip the contents of the extension, not the directory. To confirm correctness extract the modified extension, it should show extension_directory -> install.rdf file, not extension_directory -> nested directory -> install.rdf file.

This is the entire content of my RetroZilla v2.2 userChrome.css file. It's sometimes hard to see which is the active tab. Use any predefined colour and uncomment bold as desired.

/* Active tab bolded red for easier viewing */
tab[selected="true"] {
color: red !important;
#font-weight:bold
}

Link to comment
Share on other sites

Spent tedious time digging through different operating system browser ciphers and testing them in vanilla Windows 98 running RetroZilla v2.2. Tests were performed using:
https://www.ssllabs.com/ssltest/viewMyClient.html

To the best of my ability, no additional 'missed' ciphers were discovered, aside from the two recommended earlier by ClassicNick. These are also the only two green light cipher results from the SSL Labs test. As mentioned previously they help with site connections and should be added as new true booleans in RetroZilla v2.2 via about:config:

security.ssl3.ecdhe_ecdsa_aes_128_gcm_sha256
security.ssl3.ecdhe_rsa_aes_128_gcm_sha256

All other RetroZilla ciphers were reported to be weak with the following ciphers flagged as INSECURE. These insecure ciphers should be disabled (toggled false) via about:config, filter string rc4_128:

security.ssl3.ecdh_ecdsa_rc4_128_sha
security.ssl3.ecdh_rsa_rc4_128_sha
security.ssl3.ecdhe_ecdsa_rc4_128_sha
security.ssl3.ecdhe_rsa_rc4_128_sha
security.ssl3.rsa_rc4_128_md5
security.ssl3.rsa_rc4_128_sha

Edited by Wunderbar98
Link to comment
Share on other sites

Regarding site connectivity, vanilla Windows 98 with RetroZilla v2.2 and the above modifications works well, night and day difference compared to SeaMonkey v1.1.19, Firefox v2 and Dillo. For anyone wishing to run a vanilla system, recommend installing RetroZilla from the get-go as the primary browser.

Deleting the following bookmarks from my Windows 9x Web Helper script, which fail with SeaMonkey v1.1.19 but load without issue using RetroZilla v2.2:

https://sourceforge.net
https://www.autohotkey.com/docs/Tutorial.htm
https://en.wikipedia.org/wiki/Windows_98
https://wiki.debian.org/LTS
https://www.startpage.com
https://duckduckgo.com/lite
https://www.openbsd.org/
https://github.com/topics/web-browser

The following site was recently mentioned as problematic in the thread 'What does the copy Opera 12.18 into a 12.02 folder accomplish?'. With JavaScript disabled this site also loads fine in RetroZilla and scrolls to the bottom without issue:

https://www.motherjones.com

Link to comment
Share on other sites

== RetroZilla Search Engine Collection ==

= Overview =

A collection of quick coded search engines for RetroZilla builds based on SeaMonkey (not RetroZilla based on Firefox). All search engines work without activating JavaScript. Only tested on one vanilla Windows 98 system using the English language from North America (Canada), running RetroZilla v2.2 with the about:config changes recommended earlier. Search engine URLs are region neutral (eg. *.com vs *.ca).

Let me know if anyone wants help creating an additional search engine that works without JavaScript. Developers are welcome to incorporate these search engine files into new browser builds, no credit or further permission required.

= Screenshot =

https://i.postimg.cc/prV1RXN0/RZ-Search-Engines.png

= Search Engines =

Dogpile
DuckDuckGo-HTML
DuckDuckGo-Lite
Free Dictionary
FrogFind!
Google
Google Images
IMDB
Internet Archive
Mojeek
MSFN_DDG-HTML
MSFN_DDG-Lite
MSFN_Google
Startpage
Urban Dictionary
Wiby
Wikipedia
Wiktionary
Yahoo

= Installation =

The default RetroZilla search engine directory pathway is:
C:\Program Files\RetroZilla\searchplugins

Default RetroZilla search engines can either be deleted or appended DISABLE before installing the collection: dmoz (broken), jeeves (broken), google (unnecessary redirect), startpage (bloated, unnecessary redirect).

Download RetroZillaSearchEngineCollection.zip attached to this post (forum login required).

Unzip and copy all or part of the collection, including the *.png icon file associated with it's *.src search engine file, into the search engine directory.

Restart browser.

= Usage =

The Startpage search engine requires installing and utilizing the User Agent Switcher extension.

Using the context search extension mentioned earlier in this thread, select word(s) on a page, right-click and select any installed search engine from the context menu to automatically open a search in a new tab.

Or just use the usual methods to access and change search engines. Pressing F9 opens the search side bar. The default search engine also populates below the URL bar when typing a query into the bar.

= Development =

New search engines can be created using the existing *.src files as templates, adjusting as needed. Sometimes the changes are simple, or not. Similarly, decent icon files may or may not be easy to obtain.

To incorporate an icon for a custom engine, find an appropriately sized *.gif or *.png image, place it in the same search engine directory. Ensure it has the same name as the search engine file (eg. google.src uses google.png).

With major websites, often an easy way to get an icon file is to append /favicon.ico to the root URL, load the URL, right-click and save the image. The image will save as an *.ico file, rename it to a *.png (or *.gif). For example with https://duckduckgo.com/html use this URL to get the icon:
https://duckduckgo.com/favicon.ico

Edit1: All engines now have icons - thanks Drugwash.
Edit2: Added extra MSFN search engines, revised startpage.src, tweaked icon sizing, screenshot.
Edit3: Updated search engines, removed Twitter and YouTube (sites fail in RetroZilla), added Wiby.
Edit4: Updated search engines, added FrogFind! and Mojeek.

RetroZillaSearchEngineCollection.zip

Edited by Wunderbar98
Link to comment
Share on other sites

Merry Christmas to all :)

Just discovered the Firefox Legacy Collector Dump on archive.org. Here's the grim description link with some information pasted below in case the link disappears:
https://archive.org/details/Firefox_Legacy_Collector_Dump

It is unclear to me whether the maintainer is shutting down a personal site or removing the collection from archive.org, believe it's the latter. RetroZilla and old SeaMonkey utilize *.xpi extensions, query how many of these extensions might work on these old(est) browsers. Just thought some may find it useful.

The entire collection appears to be ~8 GB:

https://archive.org/download/legacycollector.org.tar.xz

The itemized list is very long and takes a long time to successfully load in RetroZilla without JavaScript. May want to open a CPU and RAM process monitor beforehand or use a little better hardware if available. To view the entire list and download individual extensions rather than the whole collection:

https://ia803100.us.archive.org/view_archive.php?archive=/16/items/Firefox_Legacy_Collector_Dump/legacycollector.org.tar

-----
Firefox Legacy Collector Dump

by
legacycollector.org


Publication date
2019-10

Language
English

To Browse the Repository: Click Here

This website is a repository for web content that has been deemed "legacy" and has been removed by their original publishers, and might otherwise be difficult or cumbersome to get.

Since starting this, end 2018, in response to Mozilla removing all legacy extensions from its add-ons site, with plans to expand to include more, similar "legacy" content, a few things have changed needing me to re-evaluate both the need for this site and my desire to run it.

Most importantly I've received several threats to my person for running legacycollector.org - not the kind of thing I signed up for.

But hey, Internet, if that is how you want to play ball, I'm out.

Considering this, I've decided to stop providing this free archive and void my plans.
This archive will be shut down in December 2019. Merry Christmas.

Mozilla has removed all "legacy extensions" from their add-ons site in November 2018, leaving many users of older "long term support" versions of the browser, as well as browser forks, dead in the water. This holds a (hopefully complete) collection of the removed extensions as they were available from addons.mozilla.org prior to this purge by the company.

Identifier
Firefox_Legacy_Collector_Dump

Scanner
Internet Archive Python library 1.8.4

Year
2019
-----

Link to comment
Share on other sites

On 12/22/2019 at 10:53 AM, Wunderbar98 said:

Thank-you for the link

Glad to be of help. :)

On 12/23/2019 at 1:43 AM, Wunderbar98 said:

[1] Missing icon

Dogpile: 32x32 PNG 16x16 PNG 16x16 ICO
MSFN: no idea!
Wiktionary: 48x48 ICO (manual resize needed)
Internet archive: 32x32 ICO (manual resize needed)
Yahoo: 32x32 ICO (manual resize needed)

I don't get the meaning for [3] and [4]. Maybe I shouldn't even try to. :)

Happy Holidays to everybody!

 

Link to comment
Share on other sites

Thanks again Drugwash.

RetroZilla Search Engine Collection post above revised, including an updated download link. All search engines now have icons, appropriately sized. The MSFN icon was obtained from https://msfn.org/board/favicon.ico.

MSFN (this site) searches via Google
The MSFN search engine uses Google's 'site:xyz' option, users that don't use Google should avoid or modify the engine.

Startpage files unaltered
RetroZilla's built-in Startpage search engine file is in this zipped collection for completeness but has not been altered in any way.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...