sanskrit-kosha / kosha

Repository to store Sanskrit koshas and scripts to process them.
25 stars 16 forks source link

Corrigenda in Dharaṇikośa text #23

Closed Harunaga-Isaacson closed 4 years ago

Harunaga-Isaacson commented 4 years ago
  1. The line अलर्को योगितः श्वा स्याच्छ्वेवतार्कपर्णपादपः ॥ has an extra syllable. It should read अलर्को योगितः श्वा स्याच्छ्वेतार्कपर्णपादपः ॥

  2. In the line कम्बू कुञ्जरशम्बूको कम्बुर्वलयशङ्खयोः ॥ कुञ्जरशम्बूको should be corrected to कुञ्जरशम्बूकौ .

  3. Similarly, in the line प्रत्युषो प्रभातवसू शैलूषौ बिल्वनर्तकौ ॥ प्रत्युषो is a mistake for प्रत्युषौ .

  4. The line तारोऽत्युच्चध्वनितरणयोः कीशमुक्ताविशुद्धयो- should read तारोऽत्युच्चध्वनितरणयोः कीशमुक्ताविशुद्ध्यो- , as also the metre shows. (Not sure if the conjunct ddhyo will display properly on all systems; in Roman transliteration, the correction consists of ddhayo --> ddhyo.)

  5. In the line ओऔशब्दौ तु होहौवत्संबुद्धयाह्वानयोः स्मृतौ। there is again what I suppose is an OCR error, ddhy having been wrongly recognized as ddhay. It should read ओऔशब्दौ तु होहौवत्संबुद्ध्याह्वानयोः स्मृतौ।

  6. In the line हे है व्यस्तो समस्तौ च हूतौ संबोधने स्मृतौ ॥ व्यस्तो should be corrected to व्यस्तौ .

I conclude that one of the repeated weaknesses of OCR at the moment is occasional mis-recognition of medial au as medial o, another being difficulty in distinguishing ddhy and ddhay. Not sure if there may be a way to check for problems of this kind automatically.

With deep appreciation and many thanks, Harunaga Isaacson

drdhaval2785 commented 4 years ago

I would say that it is not only the error of OCR. The person who proofread also had a tendency to misread triple / more conjoints. Therefore the OCR error also passed the human scrutiny.

Harunaga-Isaacson commented 4 years ago

Yes, no doubt these are difficult for the human eye as well.

drdhaval2785 commented 4 years ago

I have noticed one more error. In the line शङ्गाटकः स्याच्छैवालविशेषे [च] चतुष्पथे । the first word is a mistake for शृङ्गाटकः .