The whole article is de-indexed. That is the only way it can work - the required form of complaint is that the information is inaccurate / irrelevant / etc., i.e. the complaint is that the information should be "forgotten", not that any particular search term should not lead to it.
WRONG! The ruling is that such articles should not be returned as results when specified search terms are entered. This allows for the article to be returned in response to other search terms. Such an approach is largely in the public interest, as it means that should an article no longer be a relevant result for one person, but is still relevant for others, you can still find the article by searching for the any of the others. Take for example the following:
An article is published stating that Mr A, Mrs A, Mr B and Mr C were all arrested in connection with crime X, committed against victim Miss D.
Mr B subsequently has charges against him dropped when it became apparent that he was not involved.
The media continue to report on the case as it proceeds through the courts without mentioning Mr B.
There is no media report about the charges against Mr B being dropped.
In your interpretation of the ruling (let's call this scenario 1), searching for Mr A, Mrs A, Mr C or Miss D won't bring up the article either, despite its relevance to them.
If it is implemented according to the spirit of the ruling (and, based on my cursory and untrained reading of it, the letter of the ruling), a search for any keyword or name in the article EXCEPT for Mr B's name will bring it up, but a search for Mr B will not (we'll call this scenario 2).
And why would Mr B necessarily care if someone searching for information on Mr A turns up this article. They aren't interested in him, so for it to turn up will have no impact on him one way or another. If someone was going to be researching Mr B (for a potential employment opportunity or whatever), why would they be entering search terms like Mr A, Mrs A, Mr C, Miss D or Crime X unless they already knew enough about him that they knew about the event (and therefore are likely to know that he wasn't mentioned in later reports, or may even know that the charges were dropped.). If they do know enough about Mr Bs past to know he was tangentially involved with (for example) Mrs A, a search on Mrs A will bring up both the article where Mr B is mentioned as being arrested, and the results for the rest of the case where Mr B is NOT mentioned, allowing the searcher to infer that the charges were dropped (and if they're not able to draw such inferences, are they really the sort of people you want to be working for).
If the status quo prior to this ruling remained in place (scenario 3), you would have the situation that a search on Mr B would bring up the article linking him to the crime, but because there was no mention about the charges being dropped, additional searches would need to be made to bring up other articles pertaining to the case that omit his name, thereby providing an implication of sorts that the charges were dropped. Would most potential employers search on other terms in the article to see if the proceedings continued without Mr B's involvement?
If you put yourself in the position of Mr A, Mrs A or Mr C, you are likely to prefer scenario 1. Mr B also benefits from this scenario, but what about Miss D (or the courts, police, media or society at large), who would want to ensure that no-one else become the victims of Mr A, Mrs A or Mr C? Mr B certainly won't be a big fan of scenario 3, as it is a pretty raw deal for him and him alone. In scenario 2, Mr B is afforded a proportionate degree of protection from the ill effects of the persistence of accurate, but no longer relevant, reporting, Mr A, Mrs A and Mr C remain easily linked to a report that remains both accurate and relevant with regard to them and Miss D gets the reassurance that, whilst she remains a victim, at least the chances of others falling victim to the same perpetrators is reduced with only minimal persistent risk of harm to the reputation of Mr B. If you put yourself in the position of Miss D, the police, the courts, society at large or the media themselves, which scenario would you rather be the reality?
OK... so the last group may not really be a valid test, especially with the current ad-supported media. To them any hit is a potential source of income, so they'd rather have as many search terms hit the article as possible. If you take ad revenue/ brand recognition out of the equation, however, and look at it from a pure fairness point of view, would you not agree that the courts' decision was the fairest outcome?
The above notwithstanding, the implementation does leave something to be desired. Having the search engines handle such requests themselves is, IMHO, wrong. Each of the EU member states has, I believe, an (often independent) government department whose role is to protect the personal data of its citizens. Surely it would have made more sense to make this a responsibility of such departments, so that they can collate and vet such requests, then farm them out simultaneously to ALL search engines simultaneously, negating the need for the applicant to apply to each search engine individually, with the potential that some will accept the request, while others will not. This would ensure uniformity, and potentially negate the potential for search engines to compete on the basis of how they comply with such requests. I am not sure, however, whether the courts have such authority. It may be necessary for the EU Parliament to pass legislation to add this responsibility to those already held by these departments. If this were to happen, it would hopefully bring the added benefit of eliminating within the legislation any of the potential for misinterpretation that has resulted form this ruling (although there will always be those who, either through ignorance or deliberately being troublesome, will misinterpret it).