Volovich Mikhail. Partners
At first glance, they are completely unimportant for commercial requests. There are correlations with position in Yandex only for a small part text parameters, and even then weak. In the first three and in the third ten search results, the number of occurrences of the query in the text of the page is approximately the same - and the same can be said about the title, about headings h1-h4, etc.
But in reality, despite all this, textual factors are critically important. If we compare the TOP 30 of Yandex for commercial queries and equally relevant pages that were not included in it (from the TOP 30 of Google and Mail.ru), it turns out that the values of almost all text parameters in the TOP of Yandex are statistically significantly - and very noticeably - higher.
This most likely means that text factors work in the preliminary stages of ranking - when about a thousand results are selected, which are then further ranked in detail.
That is, if you don’t have enough occurrences of a query on your page (or individual words from the request), then your chances of getting into the Yandex TOP drop sharply.
So, for the query “chandeliers” in the TOP 30 of Yandex there are only two results where the word “chandelier” appears less than 40 times - lustron.ru with 23 occurrences and chandelier.rf with four (plus one more in the domain name). Whereas in the TOP 30 Google there are 5 pages where “chandelier” appears less than 10 times. The average number of occurrences of “chandelier” in the text of a page for the Yandex TOP 30 is 64, versus 48 for Google and 30 for pages from the Google and Mail.ru tops that were not included in the Yandex TOP.
How does the situation with text factors in Google differ from Yandex?
On Google, on the contrary, there are noticeable correlations between text factors and position - the closer to TOP 1, the higher the values. But the “input filter” is not visible, as in Yandex, and the average values of text parameters are usually somewhat lower.
But there is a lot in common. So, in both search engines, the ranking factors are stronger the further we move from the “letter” of the query. The exact form of the request, it seems, is no longer distinguished by search engines from all others. The number of individual query words scattered throughout the text is more important than how many times they appear together. Also important are the synonyms of the query words and, to an even greater extent, “additions” - those words that the search engines themselves highlight in the snippets.
Read more about this in our large analytical report on ranking factors prepared for the Optimization 2017 conference.
Let's return directly to Yandex. The search engine’s blog says that “Baden-Baden” is focused on over-optimized texts. You analyze sites under "Baden". Based on your observations, what specific features does the Yandex algorithm focus on? Why do sites fall under “Baden-Baden”?
In principle, everything is formulated quite clearly in the Yandex blog: “They contain a lot of repeating keywords and unnatural speech patterns, but few useful information. We call such texts over-optimized...”
It is easy to see that Yandex understands re-optimization in a broad way - including not only “keyword spam”, but also low naturalness and uselessness for the user.
In other words, Yandex does not like SEO texts. Moreover, he has not liked them for a long time (see mantras about sites “made for people”), but now he has learned to recognize them and decided to punish them for them.
How exactly he recognizes them is a separate and much less important question. There can be many factors, and they relate to all three components of SEO text - an abundance of keywords, unnaturalness and low usefulness. Various frequency parameters (both related to requests and not related); vocabulary typical for SEO texts; length and position of the text block; presence of hidden text; structure of the text. And this is not a complete list.
Most likely, the problem was solved in a purely practical manner. Factors that are already used by Yandex for other tasks (for example, when ranking) were used to the maximum; a number of new factors were added to them, developed specifically for this task. Machine learning ran as many times as necessary to achieve the required completeness and accuracy. Then we carried out A/B testing - and into battle.
We did not try to find out what exactly Yandex takes into account - there is clearly not enough data for this, and in the end it is not so important. We tried to learn how to identify and evaluate SEO texts. And I think we did it well. Moreover, it is important that we do not just give a summary assessment, but put everything into detail and thereby show what would be useful to correct.
What percentage of low-quality content is acceptable on a page? How much should there be to be wary of a page filter? And how long does it take for the entire site to be filtered?
Percentages do not play a special role. In order to start fighting SEO texts, Yandex had to first learn to understand well the structure of a web page - highlight navigation areas, a “showcase”, reviews, advertisements, various other blocks - and SEO text. The size of the “honey barrel” (the content part of the page) does not matter. The size of the “fly in the ointment” is important. The bigger it is, the worse it is. But it can also be relatively small. Let's say two thousand characters is definitely enough. Perhaps, apparently, a couple of paragraphs of medium size will be enough - less than a thousand characters, a hundred or so words. It's difficult to say more precisely. The fact is that there are three types of pages that are important for B.-B.:
1) those that fell under the page filter;
2) those that fell under the site filter;
3) those indicated by Yandex technical support.
And none of them provide complete clarity.
Unfortunately, we know quite little about the page filter: there were no such people among the clients of Ashmanov and Partners; users of "Turgenev" more often report a site filter, publications "in open sources", etc. is located a little. The cases known to us fall into two categories - egregious (large “terry” SEO texts, a lot of “tar”) and dubious (traffic fell shortly before the announcement of the new algorithm, so the site owner believes that this is “Baden-Baden”, but we are very we doubt it).
Much more is known about the site filter. But it is almost impossible to determine exactly which pages were taken into account when imposing sanctions, so they are also not suitable for determining thresholds. The affected sites have pages with large SEO texts that clearly deserve “Baden-Baden” - and this often includes home page website and pages of main sections. And there are also pages with very short text blocks, and without them at all. What the site was punished for is easy to understand from these data, but where Yandex draws the line is difficult.
Finally, the third type of pages are those that Yandex itself points to when responding to requests from webmasters. These pages seem to be specific, and a number of them are known, but the trouble is that they are often not very revealing. Roughly speaking, if all such (and worse) pages were under sanctions, Yandex would simply have nothing to find. One can imagine that such pages cause robots some claims, but in all cases known to us on the same sites there were much worse pages (including often the head of the site). These sites really deserve the "B.-B." - but the pages that Yandex considers necessary to show played not the first role in this.
If we talk about the market, according to your observations, how many webmasters “came to their senses” after the launch of Baden-Baden?
Surprisingly many, despite the very targeted application of sanctions so far.
Now two main competing strategies are clearly visible - play it safe or wait for the roast rooster to peck. We really don't recommend the second one.
To avoid sanctions, it is enough to remove or edit frankly unnatural texts. They are usually visible to the naked eye - but you can check with Turgenev. To get rid of the sanctions that have already been issued to you, you need to complete a quest - to guess why Plato didn’t like the page he sent, and how many more there are. Prevention is cheaper than cure.
In December you announced the Turgenev text analyzer. Tell us how and on what data you trained the analyzer? What was the sample?
There's really no machine learning there. There are algorithms and dictionaries. They were developed and debugged on the material of several hundred pages that fell under “Baden-Baden”, from about five dozen different sites, as well as many thousands of other texts, the quality of which we assessed independently.
Strictly speaking, we saw our task as learning how to effectively automatically identify over-optimized texts - and, just as important, explain what exactly is wrong with them. Machine learning would not allow this.
By the way, even in cases where a direct threat to “B.-B.” no, Turgenev is useful for webmasters and copywriters, because it helps to notice and correct unnatural repetitions, stylistic errors and other similar problems.
How does Turgenev determine that this or that content needs to be added/removed? What materials does the service compare the text with when analyzing? For example, with similar ones that are already in the database; taken from search results in real time; based on averages of all content, etc.?
We first developed the parameters responsible for the Baden-Baden risk within the Search Analytics Laboratory, and only then did the idea come up to make a publicly available service. Therefore we have two different versions“Turgenev”: the one published on turgenev.ashmanov.com, for everyone, and the “inner Turgenev”, for our clients. The internal version of the algorithm works with web pages and takes into account the context of the requests they should be found on. For example, it estimates the size of a text block and water content based on the performance of pages in the Yandex TOP for specific queries.
The public version works only with texts (this limitation made it possible to make the service free) and therefore has nothing to compare the material being tested with. He is simply assessed according to several parameters, gains penalty points if there are any for which, and ultimately according to the total points for different parameters risk is assessed. We do not recommend adding or removing content - we simply identify problems, and what to do about it is up to the author of the text or the site owner to decide.
On the “Repetitions” tab, the frequencies of words and phrases are calculated and built simple model, which evaluates the presence of “super-frequent” words. Typically, in SEO text, these are the words for which it is (re)optimized. The total number of repetitions in the text is also assessed - for this we use a parameter beloved by SEOs with the strange name “academic nausea”, which we had to reinvent, since its recipe, as it turned out, is kept secret. The thresholds for it are set quite high - penalty points start from academic nausea 10.5, i.e. only texts with a really high number of repetitions are “punished”. And the text can also get a point for the very high frequency of the conjunction “and” - this is more about stylistics in order to catch texts in which “experienced and knowledgeable doctors guarantee coding and getting rid of an addiction.”
On the “Stylists” tab, occurrences of several types of words and phrases in the text are simply counted. These are, first of all, figures of speech characteristic of bad SEO texts. For example, fragments To make sure of this choice, we suggest you view key features products or everyone knows that washing machine is an important attribute of the bathroom were sorted by us into quotes almost without a trace, and now any text in which “be convinced of this choice” is found will receive for this a certain amount of “stylistic quanta”, which after a certain threshold turn into penalty points. In addition, various stylistic errors, clericalism, excessively colloquial expressions and much more are noted. They are also taken from SEO texts, and although many of them do not in themselves make the text unnatural, if there are many of them, this becomes a problem, the severity of which is expressed in penalty points.
On the “Requests” tab, requests found in the text from a medium-sized “commercial” list (about two and a half million) are noted. If the request coverage is high, this may indicate over-optimization.
The remaining two tabs give points less often and very little at a time. “Water content” - for a low proportion of meaningful text. "Readability" - for text that contains a lot of long sentences and long words.
How much can data for the same text vary from day to day?
Usually quite a bit. We continue to improve algorithms and expand dictionaries, but the likelihood that something will change noticeably for your text is very small. A rare exception is if we used your text today to replenish dictionaries.
How accurately does Turgenev determine the risk of a site falling under Baden-Baden?
Every time we learn about a new example of a site or pages that fall under the filter, we check Turgenev’s work on them. In 95% of cases it copes - it shows a critical or high risk. In this sense, everything is fine - only we recommend checking several texts from the site (and definitely SEO texts, and not entire pages).
But this, naturally, does not mean that any text for which Turgenev showed a high risk will fall under the filter tomorrow. High scores simply indicate that the text is stilted, over-optimized, and in need of editing. There are still a lot of them in the Yandex TOP - including in the first positions, especially for information requests. Why is not a question for us.
Yandex representatives, when they are surrounded by SEO specialists and vying with each other asking why my site is under a filter, and such and such, even worse, is free, usually answer: “It’s not evening yet, everything has its time.”
We can only join this answer.
Is it true, false positives possible if texts are checked for which Turgenev is not designed. For example, many laws and other legal documents receive a lot of points for repetition and some more for style and readability. Yes, laws are also not the most natural texts, written far from best style, although not over-optimized.
“Turgenev” is designed only for commercial sites? Or is it also suitable for information portals? After all, it would seem that SEO text is both there and SEO text.
SEO texts are different, and the nature of the problems, primarily stylistic, in over-optimized informational text is somewhat different than in commercial text. From texts like Before answering the question: “Doxycycline - what are these tablets for?”, you should tell what composition this medication has you need to collect markers that are characteristic of them.
We are doing this little by little, and such texts are also being caught, although somewhat worse than commercial ones. For example, the page from which the one given in previous paragraph quote, received only 5 points, although, of course, it deserves more. But on the other hand, this text appears in the Yandex TOP in the very first position (for the query “doxycycline tablets” at the time of writing). When Yandex starts applying "Baden-Baden" to "text dumps", we will give them the attention they deserve, and they will be caught well.
We still do not know of a single confirmed example of the use of “B.-B.” to information sites. We would be very grateful if one of the readers sends them to us - as, indeed, any examples of “Baden-Baden”. The interest here is mutual - we will be happy to discuss your case with you, and it will help us improve the service for everyone.
How do you plan to develop Turgenev, in what direction?
We haven’t seriously thought about global plans yet, but locally we are working on several tabs at once. For example, we are going to radically improve the “Requests” tab - and at the same time speed up the issuance of responses (factors related to requests are now considered longer than all others).
Managing partner of the company
One of the most famous managers in Runet, a specialist in the field artificial intelligence, development software, project management. Managing partner of the company "Ashmanov and Partners". Graduated from the Faculty of Mechanics and Mathematics of Moscow State University, Candidate of Technical Sciences.
Igor Ashmanov is engaged information technology since 1983. He led the development of the Orfo spelling checker in Informatics, was a co-owner and general director of the MediaLingua company, which produced MultiLex dictionaries, and executive director of the Rambler Internet holding. During this time, he released several dozen projects, including:
- Linguistic module ORFO (program for checking spelling and style, thesaurus, hyphenation module) in the Russian version Microsoft Office, which is used by millions of people throughout Russia;
- Electronic dictionaries "MultiLex", which still remain the best professional translator tool in Russia;
- Version search engine Rambler, released in 2001, and most of the sites and services of the Rambler portal before their update in 2012.
- Spam filter “Spamtest”, protecting tens of millions of users (now sold under the brand name “Kaspersky Antispam”).
- Novoteka news search engine, which aggregates and clusters news from hundreds of sources.
- People's search "Flexum" is a service for creating thematic search engines.
Igor Ashmanov is a member of the International Union of Internet Activists “EZHE”, twice won the “Person of the Year” nomination in the ROTOR network competition (in ROTOR 2004 and in ROTOR-2006), within the framework of the “EZHE” movement there is a Gallery of prominent network figures, in which has FRI Igor Ashmanov.
Company partner
Well-known specialist in the field of software production management. Currently engaged in the development and development of the Semantic Mirror technology and a number of others. Graduated from the Faculty of Physics of Moscow State University. Teaches programming at the physics department.
In 1999–2001, Alexey worked as a manager search project at Rambler, where under his leadership it was developed and put into operation a new version search engine.
In 1995–1999, Alexey Ivanov was technical director company "MediaLingua", where he led the development of the family computer dictionaries"MultiLex" and many other projects.
In 2001–2005, he participated in the development of the Spamtest spam filter at the Ashmanov and Partners company.
In 2004–2007 - head of the Semantic Mirror project and the search engine project Ashmanov and Partners.
In 2006–2012 - development manager of the Search Engine Analyzers project.
In 2010–2011 - project manager " social search» Flexum.ru.
From 2011 to 2016, he was the project manager for Wada.vn, a search engine for the Vietnamese segment of the Internet.
In 2015–2016 - head of development of WadaMarket.com, a product aggregator for Vietnamese online and offline stores).
Company partner
Specialist in Russia in Internet technologies, development of complex websites, high loads, large volumes data. At Ashmanov and Partners he is developing a search engine.
Brought the greatest fame own projects Russian Internet Survey and the Russian Apache web server, which runs more than half of all Runet sites.
One of the best specialists in Russia in Internet technologies, development of complex websites and Internet services, high loads, large volumes of data. Graduated from the Geological Faculty of Moscow State University.
Alexey is best known for his own projects, Russian Internet Survey and the Russian Apache web server.
In 1999-2001, Alexey was the head of the Rambler’s Top100 project at the Rambler Internet holding.
Since 2001 - employee and co-owner of the company "Ashmanov and Partners".
In 2004-2006 - technical director of the company " Search technologies", author of the Novoteka news search engine and news exchange network.
Since 2008, he has been the technical director of LibRaw LLC, which develops software and development tools for digital photography.
Currently there are no projects in the company.
Alexey Tutubalin is a member of the International Union of Internet Activists "EZHE", won three times in the category "Researcher of the Year": in the network competition ROTOR 2006, in the competition ROTOR++ 2007 and in the competition ROTOR 2008, within the framework of the movement "EZHE" there is a Gallery of prominent networks figures, which contains .
Company partner
A specialist in the production of IT applications, combining the knowledge and skills of a manager, programmer and linguist. At Ashmanov and Partners he is responsible for artificial intelligence technologies.
Graduated from the Faculty of Philology of Moscow State University, Department of Applied Linguistics.
In 1996–1999, Kirill Zorkiy worked as the head of the linguistics department at MediaLingva, producing MultiLex dictionaries. Supervised the preparation of data for electronic dictionaries, developed computer morphologies of several languages.
In 1999–2001, Kirill worked as head of the department content projects at Rambler, during which time he released several dozen Rambler content projects.
Since 2001, Kirill Zorkiy has been an employee and co-owner of Ashmanov and Partners.
In 2001–2005, he was the head of the Spamtest project.
In 2005–2007 - head of the antispam department at Kaspersky Lab, head of the Kaspersky Antispam project.
Currently, he is the technical director of the Nanosemantics company.
Company partner
Candidate of Physical and Mathematical Sciences, one of the best experts in Russia in the development of technically complex software, technical audit, and IT project management.
Graduated from the Faculty of Physics of Moscow State University, Candidate of Physical and Mathematical Sciences.
In 1996–1999, Dmitry was an employee of the MediaLingua company and participated in the development of most MediaLingua projects: electronic dictionaries, search services, server applications, Internet projects, etc.
In 2000–2001, Dmitry Pashko worked at the Rambler Internet holding, first as head of the operations department and then as technical director. During this time, under his leadership, more than 30 projects were developed, tested and published. In particular, in 2000, Dmitry led the unique Internet games project “What? Where? When?”, in which teams of experts competed with the “universal mind.”
Since 2001, an employee and co-owner of Ashmanov and Partners, participated in the Spamtest, SeoRate, Semantic Mirror projects, as well as in the development of the company’s web services.
Since 2013, he has been leading the technological development of Ashmanov and Partners. Participates in subsidiary projects of the company related to Internet analysis and protecting users from unwanted information.
Company partner
Linguist, lexicographer, specialist in intelligent technologies and website usability. He has been working at Ashmanov and Partners since its inception. Currently heads the Search Analytics Laboratory.
He worked as the production editor of Kommersant (then a weekly magazine). Developed the best hyphenation algorithm for the Russian language to date. Wrote the English-Russian dictionary “Christianity”. Prepared data for MultiLex dictionaries. He was the editor-in-chief of the Rambler portal. He participated (and continues to participate) in the creation of the excellent popular science website Elementy.ru.
Since 2007, he has been working on the AnalyzeThis.ru project - these are more than 70 automatic analyzers that evaluate the quality of searches in Russian, English, Chinese and Vietnamese. He studied the phenomenon of search spam and developed methods to combat it. Managed the preparation of data and the work of linguists for the Vietnamese search engine Wada.vn and the Wada market.
In 2014-2015, she worked as director of marketing and external communications at Kribrum, where she was responsible for service development.
Since 2015, he has been managing the communications agency Ça Va Agency.
Since 2017, he has been the General Director of Ashmanov and Partners St. Petersburg.
How is it written on the Internet?Mikhail Volovich. How is it written on the Internet?
|
Doesn't it matter how
The popularity and commercial success of an Internet resource depend on “editorial” qualities - as well as on quality in general - but not primarily. Few people, when they come to the site, pay attention to such little things as spelling errors (especially if they are not too gross), inconsistencies, and poor design.
And yet the site of a reputable company, on which in the middle of the words meet extra hyphens (traces of manual hyphens) no longer looks so solid. I won’t undertake to prove it, but I personally am sure that editorial work and simply good taste pay off everywhere, including on the Internet.
Internet, Internet, Internet
Let's start with the most pressing issue: after all Internet or Internet?
In the Russian Spelling Dictionary of 1999 it was recorded Internet; Language practice is also more likely to be on his side, but the trend is clearly towards “lowering” the letter. For example, the Vedomosti newspaper (for now?) consistently writes Internet- but the Kommersant publishing house publicly announced the transition to Internet; Even earlier, some Internet resources did this.
The main arguments in favor of lowercase (“small”) letters:
The Russian language tends to turn proper names into common nouns, and with Internet this has already happened (as happened before with photocopier And diaper); Internet- nothing more than a communication environment (options: “method of broadcasting”, “information culture”), the same as telephone, TV or press.
The main arguments in favor of capital letters:
Internet still remains the name of a unique object, a proper name - the same, for example, as Mars, Interpol or Taliban. The Internet is not just a “medium” and certainly not a “method of broadcasting” (my ear refuses to perceive expressions like * report online). This is specific computer network; over the past 10 years it has grown a lot, but has retained its individuality, it has its own legislative bodies, infrastructure, etc.; it is not difficult to imagine another global computer network, parallel or alternative to the Internet.
(By the way, in English Internet- almost always the Internet, with the definite article, and there seems to be no debate about capitalization.)
If the supporters' arguments Internet seem convincing to you, or you are simply afraid of falling behind in life - write Internet. But I still believe that Internet- proper name. So I would recommend:
- Internet, undesirable * Internet;
- Net(When we're talking about about the Internet), The World Wide Web ; But world wide web Internet;
- Runet, wrong * Runet(it certainly cannot be considered a “method of translation”).
However, in compound words like internet provider, online library word Internet acts as a definition. Writing it in such cases with a capital letter is quite blunder. Another very common mistake is to write such formations in two words, without a hyphen.
- Internet cafe, Internet resource, Internet public and so on.; wrong * Internet resource, *Internet resource.
Word Internet- as well as Interpol With Taliban- should bow if standing alone. Recorded with Latin letters, Internet, Interpol, Taliban, of course, they don’t bow; but all these words have long been adopted by the Russian language, and there is no reason to write them in Latin.
- in the Internet, Internet users, Internet users; wrong * in Internet, *Internet users; undesirable * on the Internet.
Browser, tag, offline
There are many words that you come across at every turn on the Internet in a wide variety of spellings, but which until very recently were not in dictionaries.
Many of them are like myself Internet- included in the Russian Spelling Dictionary in 1999, edited by V.V. Lopatin. Many others were included in the more than one and a half thousand words that have been added to the electronic version of this dictionary since 1999.
- browser, web browser(wrong * browser, *web);
- tag, HTML tag(wrong * tag), brand, branded(wrong * brand); cache, cache(wrong * cache); hash table;
- online, online And in online mode,online; offline, offline; offshore, offshore(undesirable * online, *off-line, *offline, *offshore);
- traffic(wrong * traffic);
- player, internet player(undesirable * player, although this option is recorded in some dictionaries);
- PR(wrong * PR), PR agency(or PR agency), PR, PR man;
- realtor(wrong * realtor, *realtor; ten years ago I tried to fix it on real estate specialist, but he won); distributor(at the same time I was fighting for distributor, but also unsuccessful);
- roaming; Halloween; weekend; soundtrack;
- price list(although it is better to use a word also borrowed, but previously rooted in the Russian language price-list; wrong * price list); prime time.
The dictionary in some ways “overtakes” the established literary norm, focusing in advance on the changes that the Orthographic Commission headed by V.V. Lopatin intends to make to Russian orthography. In addition, from the current electronic version something may change for the new edition of the dictionary. Strictly speaking, it is not necessary to obey his recommendations; for example, in Kommersant they consistently write distributor And realtor, A player occurs no less frequently than player.
However, if there are “discrepancies”, it seems to me reasonable to listen to the recommendations of the most authoritative spelling reference book to date.
CD-ROM, e-mail, Java
- e-mail, e-mail", e-mail "om; Russian equivalent, better than Email or email address , not yet (but if you really want something conversational, then it’s better according to "soap" or by email, how * by email or * by email);
- CD-ROM, CD-ROM; CD, and if “in Russian”, then CD;
- Word, in Word(writing * in Word colloquially and therefore unacceptable in decent texts); Excel, about Excel, ...;
- Java applet(language name Java in Russian it is not settled, so it is better to write it in Latin).
The main thing is to be consistent
In the end, what matters is not How we write and how consistent. Word Internet you can write with capital letters, or maybe with a small one (see above). But this does not mean that it is possible to do both. We must make a decision and follow it.
When on one screen page (on a printed page, thank God, I didn’t have to) I see two news stories - one about bin Laden And al-Qaeda, and the other about Bin Laden And Al Qaeed(and one mentions a Palestinian city Ramallah, and in the other Ramallah), - I climb the wall.
It’s bad when there are periods at the end of the headings (see below) - but it’s even worse when there are sometimes periods, sometimes not.
You can include the following punctuation marks in your links, or you can not include them - but it’s better not to do this or that.
Text decoration
- No! - large unstructured volumes of text.
The maximum length of a paragraph that I (the user) is ready to read from the screen is 6-8, well, 10 lines. Then I’ll still jump: at best to the next paragraph, at worst to the next site.
It is also highly advisable to divide the text into small fragments (chapters) with separate headings.
- No! - line breaks (
). Yes! - full paragraphs (Or
).
Standard HTML leaves the site creator with very little choice: either paragraphs follow in a row and stick together, or an entire line is skipped and the text bloats. Of two evils, the first is much worse. The boundaries of paragraphs following each other without gaps are very poorly visible; the text looks like it is undivided. And if last line paragraph is long enough, the border may be completely lost.
- Yes! — numbered and unnumbered lists (bulleted lists)
You just need to make sure that you can clearly see where each element begins and ends. In this case, the list is easy to skim, which is very important when reading from the screen.
- Yes! - highlighting important places in “bold” or italics (but not in excessive quantities). No! - underscores outside links.
It’s convenient when the eye has something to grab onto; this helps you quickly scan the text and understand what you need to read in it and what you shouldn’t—a very valuable quality for the Internet. However, underlining is incorrect: it is reserved for links and any other use of it will confuse site visitors.
Capital Letters and Periods in Headings.
This headline deliberately contains an error that is very common on the Internet and very offensive in a country with such rich cultural traditions. In it (you probably noticed) all words begin with capital letters.
This is how it is customary to format headings not in Russian, but in English. (And also in English language it is customary to write with capital letters days of the week and months, names of peoples and adjectives from them. And commas in English are placed completely differently than in Russian. And the words are different. And even letters... And in German, in general, all nouns are written with capital letters.)
The Russian tradition of designing printed texts requires putting only one capital letter- at the beginning (plus proper names, etc.). Headings differ from ordinary sentences in only one way: it is not customary to put a period at the end of headings (as well as captions under photographs).
Yo
The Internet is the heir to print, not written text. Whether it's good or bad, dot the dots in printed text. e not accepted (except for dictionaries, educational literature and so on.). A chemistry book containing e, would look undignified - and just as undignified are sites where e not worth it e. (Therefore, before publishing a document in a book or on the Internet, you often have to do seemingly unnecessary and absurd work: find the marked e and replace them with e.)
Of course, books and websites where e Sometimes it’s marked, sometimes it’s not, they look even more undignified.
However, if the points above e help clarify the meaning, resolve homonymy ( did everything - did everything), they are still recommended to be installed. I usually leave e in a word All, even when confused with All hard enough.
Quotes and dashes
When working in Word, I always put an em dash and almost always - herringbone quotes, and in English text - foot quotes. (Function automatic replacement AutoCorrect is disabled for me, so I purposefully do it “by hand”). The computer taught me to format the documents I create in accordance with “book” standards, and the correct dashes and quotation marks are an integral part of them.
However, I have to make sure that in my documents and others' documents, the quotation marks are replaced with simple (", ") and the dash with a regular hyphen (-) before they are published on the Web. This is where security considerations (compatibility with other encodings and operating systems) is more important than beauty.
The point is that the dash - , Christmas tree quotes « » and paws (double “ ” and single ‘ ’ ) and some other characters (for example, number E, paragraph § , ellipsis … , ppm ‰ , midpoint , plus or minus ± ), which are in the “Windows” encoding (Windows-1251), in koi8-r and other Russian encodings, in which the user may well receive them, are missing or are in the wrong places. When outputting text in an encoding other than Windows-1251, these characters may be distorted (Cheburashkas begin to run in the text). It is possible that you see in this paragraph not the signs that I mentioned above, but some incomprehensible icons; if not, you've undoubtedly seen such icons on other sites.
Special characters like - (dash) and ‰ (ppm) do not save the situation, since there are simply no corresponding signs in koi8. The most beautiful way out of the situation is to publish texts in Windows-1251 and configure the programs so that when switching to koi, the dash is automatically replaced by a hyphen, the number sign is replaced by the letter N, and so on. This, however, requires special programming and is not always possible. Therefore, we have to reluctantly give the following recommendation.
- Before publishing text on the Internet, unless your server has special measures for the correct display of “special characters,” you must replace all types of quotation marks with simple quotes, a dash with a hyphen, and a number sign with a letter. N or the word “number”, and so on.
And once again about Rambler
- "Rambler", on Rambler; undesirable * Rambler.
Rambler chose Russian name at the beginning of 2000, it is enshrined in its official name (JSC Rambler Internet Holding). Rambler those who remember it from very ancient times continue to call Rambler Stekovsky.
- with the Internet holding "Rambler", highly undesirable * with Rambler Internet Holding!
It just so happened that a not entirely correct name was legally enshrined (to whom it doesn’t happen - look Ashmanov and Partners also written in the English manner with two capitals). This official name of Rambler can and should be used as a quotation, for example: at OJSC (company) Rambler Internet Holding. But * at Rambler Internet Holding sounds monstrous!
- Rambler's Top100 This is how it should be written. Wrong * Rambler Top100, *Rambler Top100, *Top 100, *Top100 and so on.
Notes
I quote Artemy Lebedev from the site redactor.ru: “It is noteworthy that most people writing Internet with a capital letter, they often do not decline this word and add to it the construction “international global computer network,” which is an example of blatant old-fashioned thinking.” (