Understanding prior art

Before saying "my invention is new" you have to ask: new compared with what? The answer is prior art, that is everything the world already knew before the filing (or priority) date. Understanding it is the hardest and most useful part of the work: it tells you what you can really claim and what you cannot. Here we see what it is, how to read it and what you get out of it, on an example you can look at: a telescope.

The rules quoted are those of the European Patent Convention (EPC), the European Patent Office (EPO) Guidelines for Examination and the Italian Industrial Property Code (the "Code"). The sources are listed at the end of the guide.

State of the art

"Prior art" and "state of the art" are nearly synonyms. The state of the art is everything made available to the public by written or oral description, by use, or in any other way, before the date of filing of the application (Art. 54(2) EPC; Art. 46, paragraph 2, of the Code). A single piece of information that belongs to it is called a prior-art item.

The definition is very wide (EPO Guidelines, G-IV, 1).

• Place does not matter: in Italy or abroad.
• Means and language do not matter.
• There is no age limit: a very old document counts as much as yesterday's.

The deciding date is the filing date of your application. If you claim a priority, the priority date counts for novelty (Art. 89 EPC).

What a prior-art document is

Almost anything can be one, as long as it was available to the public.

The test is accessibility. A document is accessible if, on the relevant date, people could learn its content and no duty of confidentiality restricted its use or spread (EPO Guidelines, G-IV, 1). A secret document, such as a company notebook nobody has seen, does not count. For the EPO a product on sale also counts, even if nobody is able to analyse and reproduce it (EPO Guidelines, G-IV, 2).

The description must "teach". A description is part of the state of the art only if it gives enough information for a person skilled in the art, with the common knowledge of the time, to put into practice what it teaches (EPO Guidelines, G-IV, 2).

Applications still secret. Patent applications filed before yours but published after are part of the state of the art, but only for judging novelty and not inventive step (Art. 54(3) and Art. 56 EPC; EPO Guidelines, G-IV, 5.1; for Italy, Art. 46, paragraph 3, and Art. 48 of the Code). You cannot see them while you search, because an application becomes public 18 months after filing or priority (Art. 93 EPC; Art. 53, paragraph 3, of the Code).

How to read a document against your invention

The method is simple and needs patience. You go feature by feature.

1. Write your invention as a list of features, each with its reference numeral. These are the ones a claim lists one after another.
2. Read the whole document, not only the abstract and the figures. All the content counts, not only the examples (EPO Guidelines, G-VI, 1).
3. For each feature, look for the point where it appears in the document. Note where: page, paragraph, figure.
4. Tell three cases apart: the feature is there explicitly; it is there implicitly, meaning an expert would take it for granted; it is not there.
5. Look at the overall result: which features are missing in that document, and what they do that is new.

What can be derived. A document destroys the novelty of whatever can be derived from it directly and unambiguously, including features an expert would consider implicit (EPO Guidelines, G-VI, 2). A document cannot be read as if it included well-known equivalents that it does not disclose: that is a question of obviousness, not of novelty (EPO Guidelines, G-VI, 2).

A general example. A description that names rivets destroys the novelty of "fastening means" in general, but not of a fastening by screws (EPO Guidelines, G-VI, 4). The specific destroys the general, not the other way round.

In the example. An imaginary document D1 describes a telescope with a truss, a secondary mirror, a hexapod, a top ring and an actuator, in which the actuator turns about the axis of the secondary mirror.

Only one feature is missing in D1: the offset axis. It is the difference from which the whole reasoning starts.

New and not obvious: two different questions

They are two separate judgements, and often confused.

Novelty. Does a single prior-art item, on its own, show all the features of the claim? If so, it is not new. It is not allowed to combine different documents to judge novelty, nor to combine parts of different examples in the same document, unless the document itself suggests it (EPO Guidelines, G-VI, 1).

Inventive step. Even if no document shows everything, would an expert have got there in an obvious way? Here several documents can be combined, or a document with the common knowledge of the field (EPO Guidelines, G-VII, 6).

In practice: a telescope with a single difference from D1 is "new", but may lack an inventive step, if that difference was obvious.

Problem-solution approach in plain words

The EPO assesses inventive step with a method in three steps (EPO Guidelines, G-VII, 5).

1. The starting point. The closest prior art is chosen: the one that points to the most promising path towards the invention, usually because it has a similar purpose and needs the fewest changes (G-VII, 5.1).
2. The technical problem. The differences between the invention and that document are looked at, the technical effect that follows from them is identified, and the problem the invention solves compared with the document is stated. It may differ from what the applicant had in mind (G-VII, 5.2).
3. Obviousness. The question is whether an expert, faced with that problem, would have modified the document to reach the invention. What counts is "would", not "could": not whether they were able to, but whether they would have had a reason to, expecting an advantage (G-VII, 5.3).

In the example. The starting point is D1. The difference is the offset axis. The effect is finer focusing through the useful stroke of the hexapod (28). The problem is: how to adjust the focus more finely without adding parts? Third question: would an expert, with D1 and their knowledge, have offset the axis? If a second imaginary document D2 suggested moving the axis of rotation to reduce the effect of movements, the answer might be yes. If no document suggests it, it may be no.

Who the person skilled in the art is

It is a reference figure, not a real person. It is an average professional of the field: they know what was commonly known at that date, have access to the whole state of the art and have the normal means for tests (EPO Guidelines, G-VII, 3). If the problem prompts a search for the solution in another technical field, the expert is the one of that field.

Why it matters: "obvious" means obvious to them, not to the reader, not to the inventor, and not to someone who already knows where the invention is going. The examiner must not read the prior art with hindsight (EPO Guidelines, G-VII, 5.1).

Combining several documents

For inventive step the closest document can be combined with other documents or with the common knowledge of the field (EPO Guidelines, G-VII, 6). Examiners look in particular at these points:

• do the contents of the documents make it likely that an expert would put them together? Two incompatible teachings are not normally combined in an obvious way;
• do they come from similar, neighbouring or remote technical fields?
• does one clearly refer to the other? Then combining them is normally obvious (EPO Guidelines, G-VII, 6).

Caution is needed in the opposite sense too. If the claim is a true combination, in which the features interact, it is not enough to say that each feature alone was known. If it is merely a juxtaposition, it is (EPO Guidelines, G-VII, 7).

Search report and X, Y, A citations

After filing, an office carries out a search and produces a search report: a list of the documents it considers relevant, each with a letter. For Italian applications the search report is drawn up by the European Patent Office, with a written opinion on novelty, inventive step and industrial application (Ufficio Italiano Brevetti e Marchi, UIBM, Procedimento di esame e concessione). For European applications the EPO draws it up (Art. 92 EPC).

The letters say how dangerous a document is for your claim (EPO Guidelines, B-X, 9.2).

The letters P and O always come with X, Y or A, for example "P, X" (EPO Guidelines, B-X, 9.2.3 and 9.2.4).

How to react. An X document calls for changing the claims, because on its own it hits the invention. A Y document calls for finding a difference that withstands the combination. An A document is useful for understanding the background and must not be ignored: it marks out ground already occupied.

Why a search can never prove novelty

This is the point to remember best of all. If a search finds nothing, it does not mean the invention is new.

• No search is one hundred per cent complete. Even the European search, although thorough and all-encompassing, cannot always be, because search systems are inevitably imperfect. The aim is to reduce the risk of overlooking the most relevant prior art, but it is acceptable not to find all the less relevant material (EPO Guidelines, B-III, 2.1).
• Part of the state of the art cannot be searched: applications filed but not yet published stay secret for 18 months (Art. 93 EPC).
• Something may have been made public in ways that leave no trace in a database, such as a conference, a use or a product (Art. 54(2) EPC).
• The final judgement belongs not to whoever searches, but to the office that examines: the letters in the report are opinions that examination can review (EPO Guidelines, B-III, 1.1).

The consequence is simple. A report without X or Y citations is good news, not a guarantee. In the same way, a search you do yourself before filing, however careful, proves nothing. The UIBM strongly recommends it, precisely because it can prevent investing in an application that cannot be accepted (UIBM, Cosa sapere prima del deposito), but it does not say it is enough.

Using prior art to sharpen the claims

Good prior art is not only an obstacle: it is a map. It helps with three things.

1. Deciding what goes in the preamble. In the two-part form, the preamble contains the features necessary to define the subject-matter which together already form part of the prior art (Rule 43(1)(a) EPC). The closest prior art tells you what to put there.
2. Finding the difference. The feature-by-feature table shows what is missing in the prior art. That difference goes in the characterising portion and is explained in the description with its technical effect.
3. Building the dependent claims. The features that the prior art does not show, and that you have described, are fallback positions: if claim 1 falls, claim 2 or 3 offers narrower ground (EPO Guidelines, F-IV, 3.4).

Mind the constraints. An application can be amended, but not extended beyond what it contains: an amendment cannot add subject-matter that goes beyond the application as filed (Art. 123(2) EPC). That is why every feature you might want to fall back on must be described from the start, in the description, with its effect.

The technical problem may need revisiting. The problem you wrote may not be the one the office recognises, because the prior art found can put the invention in a different perspective (EPO Guidelines, G-VII, 5.2). Better to know it beforehand.

Example telescope: recap

1. D1, imaginary, shows everything except the offset axis of the actuator (44).
2. Claim 1 is therefore new over D1, because one feature is missing (Art. 54 EPC).
3. It remains to be seen whether the difference is obvious. It depends on D2 and on the expert's knowledge.
4. The dependent claims on the distance between the axes and on the blackened bolts (50) are fallback positions.
5. No search can say that the telescope is certainly new.

Before writing yours

This page is general information, not legal advice: no reading of prior art guarantees that a patent will be granted or valid. A real application should be reviewed by a patent attorney (mandatario). In the app this corresponds to the Prior art step.

Sources

1. Art. 54 EPC, Novelty: https://www.epo.org/en/legal/epc/2020/a54.html (read on 2026-10-05)
2. Art. 56 EPC, Inventive step: https://www.epo.org/en/legal/epc/2020/a56.html (read on 2026-10-05)
3. Art. 89 EPC, Effect of priority right: https://www.epo.org/en/legal/epc/2020/a89.html (read on 2026-10-05)
4. Art. 92 EPC, Drawing up of the European search report: https://www.epo.org/en/legal/epc/2020/a92.html (read on 2026-10-05)
5. Art. 93 EPC, Publication of the European patent application: https://www.epo.org/en/legal/epc/2020/a93.html (read on 2026-10-05)
6. Rule 43 EPC: https://www.epo.org/en/legal/epc/2020/r43.html (read on 2026-10-05)
7. B-III, 1.1 Opinions in relation to the search report: https://www.epo.org/en/legal/guidelines-epc/2026/b_iii_1_1.html (read on 2026-10-05)
8. B-III, 2.1 Completeness of the search: https://www.epo.org/en/legal/guidelines-epc/2026/b_iii_2_1.html (read on 2026-10-05)
9. B-X, 9.2 and 9.2.1: https://www.epo.org/en/legal/guidelines-epc/2026/b_x_9_2.html (read on 2026-10-05)
10. B-X, 9.2.3 and 9.2.4: https://www.epo.org/en/legal/guidelines-epc/2026/b_x_9_2_3.html (read on 2026-10-05)
11. F-IV, 3.4 Independent and dependent claims: https://www.epo.org/en/legal/guidelines-epc/2026/f_iv_3_4.html (read on 2026-10-05)
12. G-IV, 1: https://www.epo.org/en/legal/guidelines-epc/2026/g_iv_1.html (read on 2026-10-05)
13. G-IV, 2 Enabling disclosure: https://www.epo.org/en/legal/guidelines-epc/2026/g_iv_2.html (read on 2026-10-05)
14. G-IV, 5.1 Art. 54(3): https://www.epo.org/en/legal/guidelines-epc/2026/g_iv_5_1.html (read on 2026-10-05)
15. G-VI, 1, 2, 4 Novelty: https://www.epo.org/en/legal/guidelines-epc/2026/g_vi_1.html (read on 2026-10-05)
16. G-VII, 3 and 4: https://www.epo.org/en/legal/guidelines-epc/2026/g_vii_3.html (read on 2026-10-05)
17. G-VII, 5, 5.1, 5.2, 5.3: https://www.epo.org/en/legal/guidelines-epc/2026/g_vii_5.html (read on 2026-10-05)
18. G-VII, 6 and 7: https://www.epo.org/en/legal/guidelines-epc/2026/g_vii_6.html (read on 2026-10-05)
19. Italian IP Code, art. 46 (Novelty): https://www.normattiva.it/uri-res/N2Ls?urn:nir:stato:decreto.legislativo:2005-02-10;30~art46 (read on 2026-10-05)
20. Art. 48 (Inventive step), art. 49 (Industrial application): https://www.normattiva.it/uri-res/N2Ls?urn:nir:stato:decreto.legislativo:2005-02-10;30~art48 (read on 2026-10-05)
21. Art. 51 (Sufficient disclosure), art. 53 (Effects of patenting): https://www.normattiva.it/uri-res/N2Ls?urn:nir:stato:decreto.legislativo:2005-02-10;30~art51 (read on 2026-10-05)
22. UIBM, What to know before filing: https://uibm.mise.gov.it/index.php/it/brevetti/brevetto-per-invenzione-industriale/deposito-di-una-domanda-di-brevetto/cosa-sapere-prima-del-deposito (read on 2026-10-05)
23. UIBM, Examination and grant procedure: https://uibm.mise.gov.it/index.php/it/brevetti/brevetto-per-invenzione-industriale/deposito-di-una-domanda-di-brevetto/procedimento-di-esame-e-concessione (read on 2026-10-05)