Wayback machine urls

i’m adding a lot of older events to mb. and to get the information i mainly us the wayback machine to go to the official website of that event (they often have time tables on there, that they often don’t have on the history page of the event website)
i would like to add the wayback machine link to the event.
do i use the “offical homepages” type or should there be a “historical page” type or something?

technically, I believe the proper practice is to only add the original URL as an “ended” URL (provided the website is no longer available, of course)

that said, I will usually add the Wayback link to the edit notes where I add the data

6 Likes

its not about the editor notes.
i would like to add it to the event url list.

1 Like

In normal MusicBrainz I put a Wayback Machine link in directly. Partly due to not everyone knowing how to find the Wayback Machine. And also sometimes you need to refer to a specific date.

1 Like

If only the ‘old official website’ gives relevant information on the event, then why not link to this ‘old official website’ via the waybackmachine. Factually it seems correct, or at least very close to correct, as it does show you the official website, just in a different time period and via the waybackmachine, but the official website nonetheless. For the end user this would be more helpful than just the link itself. And users coming from outside of MB wouldn’t look into edit notes to find this information.

For completeness sake you could add both:

  • the original link with the checkbox checked that says ‘This relationship has ended.’
  • the waybackmachine-link, that actually shows relevant information.
3 Likes

I would not link to the old festivals pages and just put it in the edit note. In general I would not add the links to the event itself. I think it is better to add the link to the Series. Usually the links you are talking about looks like this or am I wrong? Running Order Music | Wacken Open Air
This is the general link to the schedule and it always contains the actual year.

What benefit do I have from having a URL relation to an old version in wayback? The relevant data should be entered into MB and therefore you only need the wayback link in the edit note.

Not all information on a website can be captured by MusicBrainz’ data model, and in any case it’s just interesting for a lot of visitors of an event page on MusicBrainz to see a website that was made for that event way back.

Ask yourself the reverse: what would MusicBrainz loose from having a link to the official website in the Wayback Machine?

3 Likes

Yeah that’s true, but it adds some noise I think. If you are such a nerd (yes I am sometimes :smiley:) and want to look into old versions of a festivals homepage, then the chance is high, that you are aware of the wayback machine.

Often the wayback machine is pretty slow and the pages do not work properly so this leads into bad user experience.

1 Like

yeah but i think that you would only add wayback machine pages if you tested them and they are functional.

for instance, the page for pukkelpop 1997 is still functional. but if you go to 1998 then it doenst work anymore (probably because the site is either made with flash or php)
so i would only link the 1997 wayback machine page since that is functional, its no use to link unfunctional stuf in my opinion.

1 Like

I mark the relationship as ended for the original URL and add the Wayback Machine URL in the annotation like Release “Breakfast Can Wait” by Prince - MusicBrainz

3 Likes

Tickets like this one could help reduce the ‘noise’. Just voted for it :slight_smile:

Apparently waybackmachine links have been discussed extensively before.

5 Likes

When adding/editing an URL that is no longer available, the important thing to do is to set the flag “ended”. If you know when it started/stopped to be available - which can sometimes be deduced from domain name history - please set these dates (even if partial) too.

The relationship type should describe the role of the URL when it was available (“official homepage” in your example).

Ultimately, the MusicBrainz website will make use of these information to show a link to the archived page in the Wayback Machine, see the ticket mentioned above. In the meantime, using annotations (also mentioned above) is the best alternative.

Actually, the support for the Wayback Machine in MusicBrainz is actively in progress, with the current summer project by @Ashutosh_Aswal to automatically archive URLs when entered or mentioned in edit notes, see Welcome Summer of Code 2024 contributors! – MetaBrainz Blog

9 Likes

That’s nice, but like @bsammon mentions in the ticket comments, generic Wayback Machine links are a lot less useful than a link to a time stamp with actual content in it. If MusicBrainz is going to make these links automatically, I would still prefer to be able to add an exact link.

5 Likes

As @mfmeulenbelt says, sometimes it takes quite a trawl back through Wayback machine to find the right date for the page with good data on. Often Wayback machine can keep snap-shotting a site when it is already dead and throwing errors. An automatic system will need a manual over-ride.

A puzzle that is also hard to solve is how some artists have websites that changed often over the years meaning different archived dates hold data on different releases \ albums \ histories.

2 Likes

URL annotations don’t exist but still we should keep storing the original URL with its ENDED and/or dates.

Also indeed interesting snapshots, if any, are hidden in web archive search and we have to search them each time.

I think we need a new URL to URL has snapshot(s) relationship (with date*) linked to each interesting https://web.archive.org/web/<timestamp>/<path> snapshots (or URL annotation) :slight_smile:

Web service URL would be autoformatted to remove redundant https?:// from beginning of <path>.

* If possible but not possible, date of relationship should not be editable for web archive links, but should be extracted from URL timestamp and displayed in the relationship.

4 Likes

Yes, the first implementation in MusicBrainz will likely be archive links without date. But with a set ended date, we can narrow the archive link to the end year. And with taking into account the time the URL has been added to the database, we can even directly link to the closest snapshot in time that has been taken by the archiver @Ashutosh_Aswal is currently working on.

It depends on what information the person is looking for. The date range is an information that is useful on its own and that can be stored in the MusicBrainz database through the begin/end dates of URL relationships. The fact that we can generate “generic Wayback Machine links” out of it is a bonus. It is rather complementary than competing with “link to a time stamp” in the Wayback Machine.

Also, the Wayback Machine is a lot less useful when the link has not been archived in the first place and that is what the work-in-progress automated archiver is about.

It doesn’t address the issue initially mentioned in this thread, but it is the closest feature we are working on at the moment, so I thought it was worth mentioning.

One doesn’t prevent the other.

There is currently no support to help with selecting the right relationship type when adding “an exact link”. That can be an improvement for later.

4 Likes

I was referring to the annotations on the other entity type the URL would have been related too, such as an Event annotation to take the initial question as example.

Also I recommended using annotations because this fallback is currently available for entering URLs that would otherwise not be fit with the correct relationship type. Annotations are mainly meant as temporary storage of data waiting for database improvements. So if you need it, it is available, you can use it right now!

Thanks for the suggestion. Specification-wise, I would rephrase your need as “a way to say that an URL has a snapshot with the expected content and provide a link directly to it”. Implementation-wise, there are alternatives to relationships that can be easier to make and use.

3 Likes

If the automatic Wayback Machine links are an addition to the current situation, then that’s great. But it sounded like the old way with direct links to a specific snapshot was going to be deprecated, and this thread started with a discussion if these links are to be added at all.

As long as the baby isn’t thrown out with the bathwater, I’m all for the automation.