TL;DR upfront: I want to suggest we add 1) Boolean operators in the tag box filter in Advanced Search, and 2) the option/ability to save searches for easy call back and modification of a 's frequent or common searches. 622m3c

Justification for addition of "OR" operator: Tagging is -dependent and not consistent. Right now the default operator for searching tags (and genre) is "AND" (i.e., include all manga that are tagged "tag1," tag2", tag3," etc and exclude if they are not tagged with all specified). While popular tags are generally used in a straight forward manner (for instance, "Long-Haired Male Lead" does not have many variation), there are situations where multiple tags in use across the database are synonymous with each other. An example that I had posted about before is below, in relation to the redundancies and overlaps within the "Cheating/Infidelity" tag:

There is an "Adultery" tag, an "Affair/s" tag, an "Cheating/Infidelity" tag, and a stand-alone "Cheating" tag, a stand-alone "Infidelity" tag along with some more specific ones like "Cheating Boyfriend," "Cheating Girlfriend," "Cheating Wife," and "Cheating Husband," "Cheating Male Lead," "Cheating Female Lead," "Cheating Fiancé," "Cheating Fiancée," "Cheating Uke," "Cheating Seme."

This list could probably also include the "Netorare" tag. There is a level of nuance between these tags that should be maintained. However, with the way the search is set up now, a who wants to see all stories that contain this theme regardless of nuance (i.e., to review every manga that explores the topic of cheating in general), they would need to run 16 separate searches. If a less contentious or just less extensive example is needed, "Full Color" vs "Full Color Manga" vs "Colorized Version Available" is another search that would still require 3 separate searches for the same thing, depending on the tagging. Even in the context of this search, someone might be interested in any use of color in manga and the separate tags "Partially Colored" and "Accent Colors" also exist, bringing this up to five separate searches.

Assuming a is able to navigate to the tag directory and is able to assemble all the tags they want included in their search, this could be resolved with allowing the Boolean operator of "OR" in their tag search to see any manga tagged with the combination(s) that they want.

Justification for addition of "NOT" operator: In a similar vein, there are some tags s may want to exclude certain tags from their search and the operator "NOT" would be useful for this. For example, Isekai manga have become prolific and mass-produced in recent years, but there is not way of genre filter them out however they do have a somewhat fixed set of tags that could allow the vast majority to be filtered ( i.e., "Isekai" or even "Reincarnation" and more minor tags like "Transmigrated into a Book World" ). s could also filter out potentially triggering tags as well such as those related to sexual violence and/or cheating.

Justification for addition of Saved Search feature: As you can see, it takes a lot of work to look through the tag dictionary and assemble these custom searches, depending on the tag. Adding a "Save this search" feature would cut down on the work needed to run these searches the next time a is looking for new manga to read, especially if a person knows they will always want to exclude certain tags from their search every time. Likewise, since the Genre filter is also defaulted to "AND" ( so when you select "Josei" + "Romance", it will only show manga that have at the minimum the "Josei" AND "Romance" - assuming you have no exclusions, any manga with other Genres will appear as well so long as they match the minimum requirement of having "Josei" AND "Romance" ), a Saved Search feature could be a workaround to the lack of "OR" option in this field as one could pull up a Saved Search with all their tag preferences and simply change the genre filters (so if you're someone who likes Josei and Seinen, you could just uncheck "Josei" and check "Seinen" without having to re-do your tag filters).

http://i.imgur.com/LxhWm.jpg (will reformat this some day…)

Adding a specific request for the Boolean * ( asterisk/wildcard) for categories/tags in advanced find.
Right now it is only usable when searching the categories list but not in the advanced find.

To use the example flowinmyboat used all of the tags with Cheating in them could be searched with
ex: cheat This would allow for all tags with cheating anywhere in them to be included or excluded in the search

You can achieve something similar to your saved search idea if you modify the url directly on the search result page 🙂 That is the method I usually use for small changes in a search.

While I also would like it, I don't think this is a useful expenditure of time for the MU devs.

I must say, I find it baffling that you so roundly rejected minor changes to another search feature in that other thread but then come up with this, which is basically reinventing the modern search engine and all of the various coding pitfalls which come with that. Unless they copy LGPL code from somewhere, and don't bother to audit it, this is a massive time sink.

For the other discussion I agreed it would be useful but that it is easy to remove cancelled/axed results when searching in completed series by just excluding specific categories (it takes 2 seconds).

My suggestion for the ability to use * type searches solves the issue of searching for a topic like "cheating" without having to go through all the iterations one by one (and removing each one as you search, I have done this on multiple occasions when trying to find a specific story but it takes a lot of time) but also allowing the categories to still be more detailed.

I do understand the time sink for this type of work since I do large scale database design and implementation as my job. I will even volunteer to help with the work (If I can help depends on how the backend is set up).

I am not sure if this is for me, but I didn't reject the other feature? -- I agreed with them on the need for the feature, just not with some of their definitions and categorizations. This thread spawned from that one.

As for the devs, I think they already have a list of priorities, which has been mentioned in other threads, so this is unlikely to be at the top, tbh. I am not sure how this would reinvent the modern search engine...Boolean operators are very basic search functions you see on many sites "advanced search." Unless you're speaking to how MU is structured on the backend? I don't have that knowledge, so I don't know how complex this addition would be - I would assume the devs would just say, "yeah, no can do" but I am just suggesting it as a potential feature, which is the purpose on this forum. I appreciate the work people do to maintain a database of this size - I think MU is working as intended and it's clean, up-to-date, and accurate, which is saying a lot when you allow areas of input. Even tags are merged promptly when objective repeats are noted. I don't suggest this to waste time; it seems like a way to get more out of the information stored here.

@lollylopmr: Haha, yes, I do that to switch between Genres with certain tags. I was gonna try to explain it in the post, but decided against it. Another option would be to just save your theoretical tag search in a word document and copy+paste it, lol.

http://i.imgur.com/LxhWm.jpg (will reformat this some day…)

Ay you left out boolean operator searches for category filters that you mentioned in my previous thread, as well. That is as important, if not more, than booleans for tags, although the implementation should be quite similar.

It's like you're going for 2 different approaches for the same problem (unless that's what you were intentionally trying to suggest) with the suggestion of the saved search.

Most sites don't have -definable tags, and if they do, they have a whole tagging structure in place to ensure conformity thereby removing edge cases. The website you gave as an example has a rather comprehensive tagging backend... One of the most comprehensive I've seen, tbh. The hard part isn't implementing the booleans themselves; it's ensuring the datam which goes into such an open-ended tag is safe and sanitized.

I'm not saying you're suggesting anything specifically to waste time. I just don't think this is a valuable use of MU's dev time. I'm of the opinion that "good enough" is not a good enough justification for any product or service, however, I do think MU has a level of usefulness beyond what is necessary already. Pushing that further has a tendency to drop projects like MU into technical debt. What I'm saying it that it's all a balancing act between what is useful and what will end up wasting everyone's time.

Generally, what we see is that a website gets made well after the first few iterations in its early days and then has incremental upgrades thereafter. Look at Amazon or Google for example. They have billions of dollars, when was the last time their websites had a major update?

This isn't because there's nothing good that COULD be added but because technical debt (un-updated backends, outdated libraries, old spaghetti code everyone forgot how to edit) becomes a bigger problem with each major upgrade. Minor upgrades too, but major upgrades have a tendency to spiral into complete re-writes of various codebases completely unrelated to the one being modified. Depending on how MU pursued such an end, this technical debt could result in actual financial debt, which is not easy for a manga catalog to recoup.

Quote from Zackdgod

Ay you left out boolean operator searches for category filters that you mentioned in my previous thread, as well. That is as important, if not more, than booleans for tags, although the implementation should be quite similar.

It's like you're going for 2 different approaches for the same pr ...

I decided to ask for one thing at a time, haha. Since there is technically a workaround for the genre search in the most basic sense & this one is more "pressing" to me.

@Mekronid: Yeah, I get what you're saying (also I don't think I mentioned a website as an example? Or maybe this was for another ?) However, taking a step back, it sounds like you want this forum should just be limited to reporting "Bugs" then if one isn't allowed to put in their suggestions...it says in the rules "Feel free to give any suggestions in mind." I trust the MU team will be able to disregard the suggestions they find unfeasible, technically or financially, or reject suggestions they find to be without merit or overly niche. I do think this suggestion allows people to get more out of the database by allowing more complex searches, which allows them to discover things more to their tastes. MU is primarily as database and a tracking site - it's fulfilling these functions well and it has fulfilling this been since it's inception. I agree it has more utility than it's initial purpose. Based on recent announcement posts though, it sounds like the MU team is working on expanding function, so this suggestion would be in line with that goal. I don't have a list of their priorities and it's frankly not my business what they're working on; I am just a pointing out how I am using the site and identifying a potential feature that would be helpful to me and hopefully, others as well.

Also Amazon and Google have updated their websites...I am not sure what you mean by this. Of course they aren't doing to do unneeded UI or back-end updates, but that has nothing to do with their worth. You update anything as you see mission-critical or need arise. Otherwise, if what is familiar is working, why change it and confuse your base? The MU staff had a poll recently about overhauling the interface here and the overwhelming voice said not to change it (and I agree! I love the way this website looks). However, again, not sure what your this example was supposed to illustrate.

http://i.imgur.com/LxhWm.jpg (will reformat this some day…)

I also like the current look, it is efficient with a lot of data on the page without needing to do excessive scrolling.