2026-01-20: Time Zones Stretch Further Than You Think – 26 Hours Apart
Figure 1: Two extreme time zones – Baker Island (UTC-12:00) and Line Islands (UTC+14:00), adapted from International Date Line, Wikipedia.
The International Date Line divides the world into multiple time zones. Initially, the idea was to divide the globe into 24 longitudinal slices, each about 15° wide. Theoretically, this would give us 24 time zones, each with a one-hour difference, defined by an offset from a reference time, the Coordinated Universal Time (UTC). In actuality, there are more than 24 time zones. Some regions use half-hour or even quarter-hour offsets rather than full hours - for example, Nepal (UTC+5:45). In reality, time zones extend beyond 15° longitude because of political reasons, governance, economic aspects, geographical boundaries, and other conveniences. As a result, there are currently a total of 38 time zones.
Extreme Time Zones, Side by Side
The International Date Line is not a perfect straight line; rather it is a zig-zag to keep some regions within the same time zone. Figure 1 shows the part of the world’s map where we can see Baker Island and the Line Islands are geographically close to each other, but opposite to the International Date Line. This causes the two regions to be furthest apart in time, though being neighbours. Baker Island is located in the far east – world’s furthest behind time zone (UTC-12:00) and Line Islands is located in the far west – world’s furthest forward time zone (UTC+14:00). This makes Line Islands the first place on Earth to begin the next day and Baker Island the last place on Earth to end a day. Thus, the time distance between these two extreme time zones is 26 hours.
Across countries, languages, and platforms, UTC provides a single baseline for all timestamps. This is an essential choice for computer science and international standards. End-user applications localize time based on one’s device or account settings while back-end systems standardize timekeeping using UTC to eliminate ambiguity. Thus, end-user applications, like Twitter, Facebook, or email clients, typically display the time in the local timezone. However, all web archives use UTC time (i.e., 14-digit datetime) to avoid time zone confusion. Time zone offsets are primarily applied for human readability on the front-end UI and not for storing or indexing.
Estimating an Archival Time Window using Time Zone Offsets
Figure 2: A tweet posted by @randyhillier, who later deleted it.
Figure 2 shows a tweet originally posted by @randyhillier, who later deleted his tweet, and thus is no longer on the live web. However, we can search in the web archives to see if the tweet was archived. To search for the tweet in the web archives, we need to make a reasonable estimate of the earliest possible time the tweet could have been archived.
The timestamp (9:41 PM Feb 19, 2022) shown in the screenshot in Figure 1 reflects the timezone of the person who took the screenshot. However, the timezone of the user that took the screenshot (the "viewer") is not necessarily the same timezone as the original poster (or "author", in this case @randyhillier). Let us consider a scenario where the viewer and the author are at two extreme time zones. The timestamp on the screenshot "Feb 19, 2022 9:41 PM" is in the time zone of the viewer who took the screenshot. If the viewer is in the UTC+14 time zone and the author is in the UTC-12 time zone, then the author’s timestamp of the tweet would be 26 hours behind the viewer’s timestamp (Figure 3 - top). Similarly, if the time zones are swapped between viewer and author, the author’s timestamp of the tweet would be 26 hours before the viewer’s timestamp (Figure 3 - bottom).
Figure 3: Determining a reasonable estimate of the earliest archival time of a tweet to search in the web archives.
As a result, the archival time window for the alleged author’s tweet could be estimated as ±26 hours from the screenshot timestamp. Therefore, it would be reasonable to use a left-hand boundary of 26 hours before the screenshot timestamp for searching in the web archive. Since a URL cannot be archived before it exists, we can eliminate all URLs archived earlier than the timestamp in the screenshot, minus 26 hours. We do not limit the right-hand boundary since early posts may be archived at later times.
The following curl command shows the total number of tweets archived when we query the Wayback Machine’s CDX API using only a Twitter handle. The CDX response results in 36,772 tweets for @randyhillier's status URLs.
Figure 4: Using the CDX API to retrieve the total number of tweets archived for a user using the Twitter handle.
The next curl command shows the total number of tweets archived when we query the Wayback Machine’s CDX API using a left-hand boundary of 26 hours before the screenshot timestamp along with the Twitter handle. The CDX response results in 16,258 tweets for @randyhillier's status URLs which were archived 26 hours before (7:41 PM Feb 18, 2022) the screenshot timestamp (9:41 PM Feb 19, 2022).
Figure 5: Using the CDX API with a left-hand boundary along with the Twitter handle to limit the search space.
Thus, the total number of archived tweets required to individually dereference is reduced by over half when a left-hand boundary is used along with the Twitter handle.
Summary
It is natural to assume the time difference across the globe is a maximum of 24 hours. However, for geopolitical reasons the global time range is actually 26 hours. We use the range between two extreme points on Earth to estimate an archival time window for a tweet in the screenshot, thereby often allowing us to reduce the number of archived tweets that have to be individually dereferenced.
— Tarannum Zaki (@tarannum_zaki)
Comments
Post a Comment