Preamble

Some few years ago, Rick Thomas started collecting the "interesting" messages to the comp.protocols.time.ntp newsgroup. Over time, the collection got pretty large. He did some re-formatting and editing, and published it as the NTP "pseudo-FAQ". It wasn't pretty, but it did the job. Rick tried to publish updates every month or so, but then his life got busy and the FAQ went for a couple of years without any updates. There's been a lot happening in the world of time-keeping in those years, and that old "pseudo-FAQ" is now badly out of date.

Largely out of habit Rick has kept up the collection, but he's not had the time to edit it into a useful form, and it's become clear that he's not going to have the time in the foreseeable future. Fortunately, Ulrich Windl recently offered to take the raw collection in hand and make something useful out of it. What you see here is the result.

A short explanation of the result (written by Ulrich Windl):

(In case of comments for this section, please quote the following information in your report: $RCSfile: preamble.html,v $ $Revision: 1.1 $)

To make the huge file with all the articles more handy, it is split into several parts automatically. The hyper-linking program (that has been written for this collection of articles) extracts authors, subjects, message IDs, and selectable keywords from the articles to build indices or other types of hyper-links for these categories.

Indices referring to these articles are possibly also split and (once again) indexed by the main document (You are currently visiting it).

That was necessary because the collection of articles was a rather large file that several HTML browsers could not handle correctly nor efficiently.

To make the articles (containing tabular output or other ASCII artwork) most useful, the converter tries to preserve the original layout wherever possible.

For convenience you can not only jump from the indices to the articles, but also back to the indices (as well as from part to part, plus some other wild things):

Selecting the subject of an article, you'll reach the index of all articles (See it) with the same subject (disregarding common reply prefixes).
1. Within that part of the index the articles are listed without any sorting (that is ordered by arrival time). For convenience the article's author is shown for each item (duplicates are possible).
2. Each list item (author) is hyper-linked to the corresponding article.
Following the hyper-link of the author of an article, you'll reach the index of authors (See it). Optional hyper-links in the body of an article can refer to other (or even the same) authors. The index names all articles sent by that author. For convenience the subject is shown for each article.
1. For each author in the index there are at least two links: The first one (author's name or EMail address) is a mailto-link, while the following links (subjects) refer to corresponding articles.
For reasons of performance and file size, the mechanism is different for the optional index of keywords (See it). In the default case keywords are neither highlighted nor linked inside the articles as most browsers have a search function (If you need that functionality).
Nevertheless I have added a new header field named ``X-Keywords'' that lists matching keywords in the body of the article (Keywords are separated by one space form each other). Remember that you can get back to the last link visited using the browser's `Back' function.
If enabled, matching keywords can be highlighted and hyper-linked back to the index of keywords (similar to authors). That might help you to locate related articles.
The selection of keywords is a difficult and manual task.
- Neither the least, nor the most frequent words are good candidates. Instead it depends on the subject discussed. I decided to require at least three letters for a keyword, and I did not want more hits than 50 articles per keyword (So you probably won't find "help" or "error" in the index).
- I only wanted to have few keywords to limit the size of the index. Remember that even though the selection of keywords is a manual task, the assignment of articles to keywords is done automatically. Therefore you might find a hit in someone's signature or quotation.
Now that Performance since version 0.36 of the hyper-linking program has been dramatically improved, let's explain and enjoy the additional goodies:
1. The keyword index (See it) lists both, subject and author of the corresponding article. For no other reason but convenience, the subject is hyper-linked with the article, while the author is linked with the index of authors.
There is another optional hyper-link for Message-IDs: If the header of an article has `References:' or `In-Reply-To:' fields, and that message being referred to has been found, the hyper-link will bring you to that article.
For reasons of performance, no hyper-links for message IDs are currently created within the body of an article.
If enabled, additional hyperlinks referring to common URLs (ftp:, http:, mailto:) are made active and they bring you to the indicated location.
For convenience of sequential reading there are links at the end of each file that link to the next part (or to the main part). A corresponding mechanism exists at the start of the file.
There are some new optional hyperlinks at the end of some header fields, all appearing as ``[-/+]''. These links let you jump backwards and forwards within one category:
1. At the ``Date:'' header you can go to the previous or next article, assuming the articles were collected in order.
2. At the Header specifiying the author, usually ``From:'', you can go to the previous or next article written by the same author.
3. At the header field ``Subject:'' you can select the previous or next article about the same subject.
4. At each keyword in the ``X-Keywords:'' header field you can go to the previous or next article containing the selected keyword.

These were the main features that the resulting HTML file have. Even if you feel uncomfortable at the beginning, I hope you'll like it eventually.

Currently the hyper-linking program can not incrementally add new articles; everything has to be rebuilt from scratch. Therefore I'll appreciate any offline collections of interesting articles. All programming and formatting activity happened in my spare time.

[End of explanation]

Note, This is a "non-fat" FAQ. We include stuff from the newsgroup that we think is interesting. We include it as it comes to us, warts and all. Rick doesn't usually write things himself because he doesn't have the time, though readers of the newsgroup will recognize Ulrich as a frequent contributer.

For legal reasons we have to emphasize that neither this program nor any of the people involved makes any attempt to evaluate the individual articles for correctness, or in any other way except the purely subjective criterion of being "interesting".

Because of the un-edited format, you should read all relevant items before you act on anything you read here. Frequently, later items contain corrections to earlier items.

Please post updates to the newsgroup.

This was last updated $Date: 1998/03/30 20:24:39 $

Rick Thomas, LCSR Computing Facility, CoRE Building
Rutgers University, Department of Computer Science
Busch Campus, Piscataway, NJ 08855-0879, USA
Internet: rbthomas@cs.rutgers.edu

Ulrich Windl, Rechenzentrum DV-med
Klinikum der Universität Regensburg
Franz-Josef-Strauß-Allee 11, 93042 Regensburg, Germany
Internet: Ulrich.Windl@rz.uni-regensburg.de

Standard disclaimers apply. All messages quoted here are the opinions of their respective authors, and do not necessarily represent the opinions (if any) of their respective employers, or of the compilers of this FAQ. If you act on the advice contained herein you do so at your own risk. Nobody involved makes any warranty.

Preamble

Contents