Armenian PowerSpell secures 97% of language coverage

Interview with the author of Armenian PowerSpell Program Arman Boshyan and Director of “Sahak Partev-Papyan” Ltd. Hayk Papyan to Mediamax Agency and Itel.am portal

- How long have you been dealing with production of Armenian PowerSpell program? The full name of the program is Armenian PowerSpell Extended Edition 5.2. So, it turns out that there have been previous versions as well? 

A. Boshyan: The prototype of Armenian PowerSpell – ArmOrfo, was created 9 years ago and was the first program, which checks orthography of the Armenian language. Such programs are usually called spell checkers. ArmOrfo possessed quite limited capacities: the dictionary contained 60 000 words. Nevertheless, the program enjoyed quite significant demand, since there were no alternatives to it at the time. The program was quite poorly protected from illegal distribution, the fact becoming an obstacle to its development. After the issue of the new Microsoft Windows operational system, illegal distribution of the program stopped, since ArmOrfo did not operate with the new Windows.  

Thus, the program stopped developing because there was no necessary funding. In 2005, I made a decision to continue funding at the expense of my own resources and give a new birth to the program, calling it Armenian PowerSpell. I involved linguists and translators in the work. We decided to create a program, which would take into account the rules of Armenian orthography, the simple grammar, and, most importantly, which would work basing on phonation.

Armenian PowerSpell is the only spell checker in the world today, which checks how the word sounds, and gives the corresponding orthographic suggestions. Thus, the program takes into account both the radical, and its case and form of use. 

Illegal distribution of PowerSpell 2005 hindered its development. Armenian PowerSpell 2009 is my third effort, which I again realize at the expense of my own means, but already taking into account the efforts of illegal distribution and taking up all the necessary preventive measures.

The program was put for sale on March 1, 2009. It secures 97% coverage of the Armenian language.

- What does “97% coverage of Armenian language” mean? 

A. Boshyan: This means that the program recognizes 97% of words used in Armenian language (approximately 550 000 words and word combinations), and, naturally, it corrects the mistakes in those words. By the way, the user may individually add words, which are absent in the dictionary, to the program.

97% coverage fully meets international standards. Analogous programs, which check the orthography of Russian, Spanish and French languages, secure 95-96% language coverage. Taking into account the fact that the Armenian language is one of the most difficult languages in the world, in order to reach 97% coverage, we needed 5 years’ hard work of 20 specialists.

The program checks Armenian orthography, simple grammar and punctuation, sequences, difficult for perception, certain syntactical mistakes, use of auxiliary verbs. It also analyses foreign words and suggests Armenian variants.

Composing the list of foreign words, included in the program, we have studied newspapers and internet forums, and singled out 400 non-Armenian words, which are widely used in the society.

I hope that in 2010, Armenian PowerSpell will become a full-fledged program for grammar check.

- Familiarizing with the program, we noticed that it only suggests corrections, while, for instance, analogous Russian and English programs automatically correct certain words, write with a capital letter after a period. Why doesn’t the Armenian PowerSpell have such a convenient opportunity?

A. Boshyan: Armenian PowerSpell protects the right of a person to express oneself the way one wants. The program will not correct any mistake without the corresponding command, taking into account the possibility that the user makes the mistake consciously.

- The program description notes that the orthography is checked by two methods: simple and strict. Do those methods function simultaneously or does the user individually decide, by which method to check?

A. Boshyan: The user decides. The simple method corrects in case, for instance, the user has typed an unnecessary or a wrong letter, and the strict method checks the cases of the words and if the user enters a foreign word, which has its analogue in Armenian.

The simple method works speedily and is convenient for everyday letter exchange, while the strict variant functions slowly, but secures 99.99% orthographical accuracy. This method is most expedient for publishing houses. The strict method of check is recommended for use also while composing important documents. 

- Does the program work only with Armenian Unicode, or does it also recognize ANSI encoding, which is used today by the majority of users?

A. Boshyan: The program works with all encoding types of Armenian language. 

- Typing by ANSI, letters often independently change. Does Armenian PowerSpell eliminate this shortcoming?

A. Boshyan: Yes, it does. Moreover, the program also takes into account cases, when the user, typing in Armenian, by mistake uses Latin letters, which correspond to the Armenian ones by their form.

- How is the program distributed? How many organization and individuals have already purchased it, and do you plan any advertisement campaigns to make it more popular?

A. Papyan: “Areg” Inquiry Office, which belongs to “Sahak Partev-Papyan” Company, deals with distribution of the program. 

I cannot but note the symbolic chance that a company, which contains the name of Catholicos Sahak Partev in its name, deals with urgent tasks of the Armenian language. It is known that Mesrop Mashtots created the Armenian alphabet by the very order of Sahak Partev.

The program is distributed by means of www.powerspell.am and www.soft.areg.biz websites, which are quite well indexed by search systems. Besides, we have direct contact with thousands of organizations, the fact allowing us realizing quite efficient market research and distribution of the program.

As of today, Armenian PowerSpell has already 300 subscribers. Besides, the program is free for 7 days’ use. By means of our central server, we track the frequency of the program use and can establish that in the course of 2 months the program has checked 160 000 pages and corrected 800 000 mistakes.

We constantly receive positive responses from users. In reality, the price for the program, which makes 9000 AMD annually, is not comparable to its benefit.

By distributing Armenian PowerSpell, we strive to import the culture of using licensed programs to the Armenian society. 

The program is constantly updated, and the owner of its licensed version may independently choose the most convenient method to update it.

- Is the program distributed only in Armenia, or also abroad?

A. Papyan: At the moment, our target market is Armenia. Unfortunately, in the Armenian Diaspora the practice of typing in Armenia is underdeveloped, and there is almost no demand for such a program.

- In the nearest future, Armenian version of Microsoft Office 2007 will appear. Is there the possibility to sell Microsoft Office and Armenian PowerSpell in a single program package?

A. Boshyan: We think about that. Usually, weaker versions of spell checkers are spread with Microsoft Office, and the professional programs are sold separately. Probably, we will do it that way.

- We know that besides Armenian PowerSpell you also elaborate an electronic dictionary Armenian PowerDictionary. Can we suppose that this dictionary will provide the Armenian language speaking users with opportunities, such as ones, for instance, provided by ABBYY Lingvo dictionary to Russian language speaking users?

A. Boshyan: In a few months, Armenian PowerDictionary 2009 will come out. It includes 200 000 words. The dictionary will be in three languages – Armenian-Russian-English. I assure you, it will provide the Armenian language speaking user with far wider opportunities than, let’s say, the Russian language speaker is provided by Lingvo.