"Problem" word filter can be circumvented in a search engine compatible way

Question

This is not about the merits of the filter. But it exists, and its implementation is so limited it is ineffective, with good, easy workarounds. I consider this to be a bug.

You might feel different and consider this to be a feature and can now go crazy fixing the more useful occurrences of problem in titles.

If you want to voice your opinion on the filter itself, this is not the topic for it. Please see the following topics:

• Remove blacklist filter for `problem`
• Why can't we use the word "problem" in titles?
• The Efficacy Of Stack Overflow's Question Title Filter
• Filter titles that only contain tags and/or "filler" words
• Let users with sufficient reputation use "problem" in titles
• The title word filter is one of the worst ideas ever implemented on SO
• Block, but automatically flag "Problem" attempts in the title as "low-quality"

Introduction and previous attempts

It is easily possible to circumvent the problem title filter. Now, there were horrible workarounds like the leet pr0blem, or proble and problam as previously mentioned on MSO.

IMO, the most elegant solution so far is @slhck's, which uses lookalike characters from other character sets. They might even work with Google. This solution has two drawbacks though: The workaround is obvious through the ASCII-only search engine friendly question URL (which contains e.g. problm instead of problem), and at least with Safari's search function, searching for problem does no longer find the title of the topic.

A better workaround

I was able to edit this post in a way that works with both Google and in-browser search.

Edit diff

Google snippet

 In-browser search

Description

I inserted a Unicode character after each occurrence of problem that is considered to be part of the word: Zero-width joiner U+200D. It might also work with the zero-width non-joiner and zero-width space. I didn't test those.

The site search itself fails to find the topic, since the character is part of the problem term. Both my browser and Google ignore it though.

How do I insert this character?

Open /System/Library/InputMethods/CharacterPalette.app on a Mac (e.g. via Edit » Special Characters… in pretty much any application), and search for 200d. In the second list will be an invisible item whose outline becomes visible once selected. Drag and drop onto an editing text field.

Answer 1

The filter is intended only to encourage users to re-think their title. It is not intended to fully stop the use of the word "problem" in the title.

Were we to fanatically pursue filter perfection in this regard, we would be working endlessly because we, as programmers, are pretty good at finding corner cases. It's not worth it, since the only time this filter is meant to take action is when a well-meaning user uses the word in the title.

We are not targeting abusive users with this filter - we have a variety of other means to encourage them to use the site better. balpha sums it up pretty well:

There are many work-arounds for certain limitations. Sometimes we fix them, sometimes we say "meh". If the restriction prevents a couple of [bad posts], it has fulfilled its purpose. ... When you allow users to post stuff, there's always going to be someone who uses the magic of unicode to do smart and/or funny stuff.

As long as it's something that can be looked over, that's fine. If it happens too much, we'll prevent it, but it'll always be a game of cat and mouse – sooner or later someone finds something else.

If you come across a question which uses the word "problem" inappropriately and appears to be a deliberate abuse of the system, flag it as such. Otherwise, let the filter gently remind new users of their obligation to provide a descriptive title, and ignore users who post good titles that use a workaround so they can use the language they desire, even if it's discouraged.

Answer 2

I think it's reasonable of filter designers to expect that most users capable of exploiting advanced trick like that are also capable of finding better alternatives.

Using your example, as a regular (non-moderator) user, I would prefer other, simpler and more straightforward way to handle this problem than messing with special characters:

In my experience it typically happened that question text provides better replacement for "problem‍" - including SEO perspective.

above, "problem‍" is written using mentioned trick "problem‍"